Journal of Evidence-Based Medicine ISSN 1756-5391

REVIEW

Impact of pay-for-performance on management of diabetes: a systematic review Jin Huang1,∗ , Senlin Yin1,∗ , Yifei Lin1,∗ , Qian Jiang2 , Yazhou He1 and Liang Du2 1 2

West China School of Medicine, West China Hospital, Sichuan University, Chengdu, China Chinese Evidence-Based Medicine Center, West China Hospital, Sichuan University, Chengdu, China

Keywords Diabetes; meta-analysis; pay-for-performance (P4P); systematic review. Correspondence Liang Du, Evidence-Based Medicine, Chinese Evidence-Based Medicine Center, West China Hospital, Sichuan University, No. 37 GuoXue Xiang, Chengdu 610041, China. Tel: +86–028–86426295; Fax: +86–028–85422052; Email: [email protected]

Jin Huang, Senlin Yin, and Yifei Lin contributed equally to this work Finacial disclosure This research was funded by “The China-U.S. Center on Medical Professionalism” (Project No., PUHSC-MPC1001). The funding organization did not involve in the design and conduct of the study, analysis, or interpretation of the data, or in the preparation of and decision to submit the manuscript. Declare of contribution All authors of this paper have directly participated in the planning, execution, or analysis of the study. All authors of this paper have read and approved the final version submitted.

Abstract Objectives: To review and synthesize published evidence of pay-for-performance (P4P) effects on management of diabetes. Methods: Databases including Ovid MEDLINE, EMbase, PubMed, The Cochrane Library (Issue 3, 2012) were comprehensively searched for the effects of P4P programs in terms of patient outcomes and physician behaviors. Studies covering detailed data were included and synthesized. The quality of the body of evidence for each quality indicator was determined using Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system. Results: Among 742 identified articles, 12 interrupted time series studies, 7 controlled before-after studies, and 2 cross-sectional studies were included. Additionally, 12 studies were further included for quantitative analysis. Results of metaanalysis showed that P4P produced generally positive effects in most indicators (eg, patients with records of total cholesterol or blood pressure). However, these results were inconsistent. The percentage of patients with HbA1c ≤ 7% or 53 mmol/mol showed a pooled odds ratio of 0.98 in patients, but a pooled mean difference of 19.71% in the physician groups. The odds ratios of receiving tests/reaching an outcome level were also diverse in patients (odds ratios ranged from 0.98 to 3.32). Besides, process indicators had higher rates of improvement than outcome indicators. Conclusions: P4P programs have variable impacts on patient outcomes of diabetes as well as physician behaviors, with various effects from negligible to strongly beneficial. Considering the low quality of the included studies, this conclusion should be cautiously interpreted.

Conflict of interest The authors have no conflict of interest. Received 26 June 2013; accepted for publication 15 July 2013. doi: 10.1111/jebm.12052

C 2013 Wiley Publishing Asia Pty Ltd and Chinese Cochrane Center, West China Hospital of Sichuan University JEBM 6 (2013) 173–184 

173

Impact of pay-for-performance

J. Huang et al.

Introduction Diabetes mellitus is a commonly seen disease which intensively impairs patients’ quality of life. And it is also one of the major financial burdens of healthcare system worldwide. Efforts have been made to achieve better management of diabetes. It is suggested that effective management of diabetes depends on not only right medication, but also on proper compliance of patient and physician behavior. Payfor-performance (P4P) is a financial incentive that directly relates a proportion of the remuneration of providers to the achieved result on quality indicators (1). Since it is clear that physician behavior is affected by financial incentives (2), P4P may maximize the effects of existing therapies and medications, and take management of diabetes to a whole new level by changing physician behavior. In recent years, P4P has been widely adopted, and increasingly recognized in intervening management of chronic diseases. A growing number of studies have reported that P4P effects on management of diabetes. However, due to the heterogeneity of P4P settings as well as demographic differences, results are generally inconsistent and even controversial. Previous reviews either failed to focus on P4P in diabetes field or insufficiently synthesized the results (1,3,4). This systematic review presents results based on comprehensive search in electronic databases. Evidence of P4P affecting management of diabetes was reviewed and synthesized using meta-analysis. The study aims to: (a) explore whether P4P positively influence quality indicators of diabetes mellitus and the size of the effects; and (b) evaluate the quality of the body of evidence for each relevant indicator using Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system.

Methods Inclusion and exclusion criteria (a) Inclusion criteria

r r r r

Participants: Primary care physicians, physician groups, general practitioners involving the treatment of patients with diabetes; patients with diabetes being treated. Intervention: P4P programs in which the whole or part of physicians’ income is related to the quality of medical health service that they provided. Control: Studies should set either before-after control groups, or control groups with no intervention. Outcome: Studies should report quality indicators with concrete data concerning direct or indirect biochemical test/physical examination results of patients (eg, HbA1c level), or changes in clinical behavior of physicians includ-

174

r r

ing performing tests, prescribing, admission, or referral, etc. Both X ± SD and group counts were acceptable. Study design: Randomized controlled trials (RCTs), controlled before-after studies (CBAs), or interrupted time series (ITS). Language: Only studies published in English were included.

(b) Exclusion criteria

r r r

Perspectives with no original data. Articles introducing specific P4P settings. Outdated reports of continuous studies (only the newest reports were included).

Search strategy Electronic search was conducted in the first week of February 2012. Database were searched including EMbase, Ovid MEDLINE (February, 1946 to 2012), The Cochrane Library (Issue 3, 2012), and PubMed, Ovid In-Process & Other NonIndexed Citations (updated in 2 February, 2012). Similar search strategies were applied to different databases. “P4P,” “general practice,” and “diabetes mellitus” were selected as the basis for index terms (eg, MeSH terms) and free text terms. The PubMed search strategy was then translated into other databases, using controlled vocabulary. Detailed search strategies for each database were presented in Box 1.

Box 1 Search strategies in different databases (n) EMbase #1 'pay for performance' (1,240) #2 'p4p'(280) #3 'reimbursement'/exp AND 'performance' (1,987) #4 1 OR 2 OR 3 (2,912) #5 'general practitioner'/exp OR 'general practitioner' (60,946) #6 'general practice'/exp OR 'general practice' (94,319) #7 5 OR 6 (138,508) #8 4 AND 7 (177) Ovid MEDLINE #1 p4p {including related terms} (250) #2 pay-for-performance {including related terms} (979) #3 reimbursement, incentive/ (2,358) #4 performance {including related terms} (20,074) #5 #3 AND #4 (199) #6 #1 OR #2 OR #5 (1,119) #7 general practice.mp. OR family practice/ OR general practice/ (73,437) #8 general practitioner.mp. OR general practitioners/ (12,645) #9 #7 OR #8 (80,339) #10 #6 AND #9 (57) The Cochrane Library #1 'pay for performance' (0) #2 (reimbursement) (1,518) #3 (performance) (35,932) #4 (#2 AND #3) (168) #5 (p4p) (9) #6 (#4 OR #5) (168) #7 (general practitioner) (5,231) #8 (general practice) (14,618) #9 (#7 OR #8) (16,081) #10 (#6 AND #9) (83)

C 2013 Wiley Publishing Asia Pty Ltd and Chinese Cochrane Center, West China Hospital of Sichuan University JEBM 6 (2013) 173–184 

J. Huang et al.

PubMed #1 "reimbursement, incentive"[Mesh] (2,355) #2 "performance" (484,443) #3 "pay for performance" (1,021) #4 p4p (257) #5 #1 AND #2 (668) #6 #3 OR #4 OR #5 (1,341) #7 "general practitioners"[Mesh] (655) #8 general practitioner (41,365) #9 "general practice"[Mesh] (58,903) #10 general practice (124,980) #11 #7 OR #8 OR #9 OR #10 (142,315) #12 #6 AND #11 (111) Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations #1 p4p {including related terms} (12) #2 pay-for-performance {including related terms} (51) #3 reimbursement, incentive/ (0) #4 performance {including related terms} (27,078) #5 #3 AND #4 (0) #6 #1 OR #2 OR #5 (52) #7 general practice.mp. OR Family Practice/ OR General Practice/ (881) #8 general practitioner.mp. OR General Practitioners/ (533) #9 #7 OR #8 (1,352) #10 #6 AND #9 (2) DM supplement search PubMed #1 "reimbursement, incentive"[Mesh](2,355) #2 pay for performance (4,004) #3 p4p (257) #4 #1 OR #2 OR #3 (4,041) #5 "diabetes mellitus"[Mesh] (273,670) #6 #4 AND #5 (86) Ovid MEDLINE 1946-present #1 p4p {including related terms} (254) #2 pay-for-performance {including related terms} (1,005) #3 reimbursement, incentive/ (2,338) #4 performance {including related terms} (23,017) #5 #3 AND #4 (231) #6 #1 OR #2 OR #5 (1,143)

#7 diabetes mellitus/ (83,957) #8 #6 AND #7 (35) EMbase #1 'pay for performance' (1,240) #2 'p4p'(280) #3 'reimbursement'/exp AND 'performance' (1,987) #4 #1 OR #2 OR #3 (2,912) #5 'diabetes mellitus'/exp OR 'diabetes mellitus' (501,841) #6 #4 AND #5 (184) The Cochrane Library #1 'pay for performance' (0) #2 (reim bursement) (1,518) #3 (performance) (35,932) #4 (#2 AND #3) (168) #5 (p4p) (9) #6 (#4 OR #5) (169) #7 MeSH descriptor Diabetes Mellitus explode all trees (13,292) #8 (#6 AND #7) (7)

Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations #1 p4p {including related terms} (12) #2 pay-for-performance {including related terms} (51) #3 reimbursement, incentive/ (0) #4 performance {including related terms} (27,078) #5 #3 AND #4 (0) #6 #1 OR #2 OR #5 (52) #7 diabetes mellitus/ (0) #8 #6 AND #7 (0)

Impact of pay-for-performance

Study selection and data extraction Search results obtained from five literature databases were input to EndNote X4. two reviewers independently screened the studies by titles and abstracts according to the predefined inclusion and exclusion criteria. Full texts were further retrieved and examined. The included studies were categorized according to study design. Baseline characteristics were then extracted. The Quality and Outcomes Framework (QOF) indicators (5) in diabetes domain were applied to sort reported quality indicators into closest QOF indicators. The updated American Diabetes Association (ADA) guideline for standards of medical care in diabetes (6) was then adopted to divide the quality indicators into several domains. The indicators were further divided into process indicators (clinical behavior of physicians, such as ordering a test) and outcome indicators (biochemical test/physical examination/history taking result of patients, such blood pressure level). The indicators were listed in Table 1.

Data synthesis and statistical analysis Results were then extracted and input to RevMan 5.0 for statistical analysis. If data pooling was possible for an outcome, synthesized results would be presented using odds ratio (OR) or mean difference (MD) with 95% Confidence Interval (CI) and P value. If an outcome was reported in less than two studies or studies with different design/participants, meta-analysis would not be performed. Instead, descriptive analysis would be performed. The results of included studies were sorted by the kind of participants: (a) studies reporting data directly from patients (eg, percentage of patients with blood pressure records was 70%); (b) studies reporting data directly from the physicians/physician groups (eg, the average percentage of patients with blood pressure records from three clinics was 72%, with standard deviation of 10%). Standard Chi2 test was applied for heterogeneity in the comparison of different studies that reported control/event count, and t test was used to analyze studies that reported mean value and standard difference (SD). Pooled OR and P value were calculated using RevMan 5.0.

Evaluation of evidence Methodological quality of the included studies was evaluated using several appraisal tools. Two reviewers independently applied Quality Index (QI) (7) to evaluate the quality of RCT, CBA, and ITS. Also, EPOC 2009 risk of bias guideline (8) was used to assess risk of bias within and among studies. GRADE (9) system was adopted to level the quality of the body of evidence for each quality indicator.

C 2013 Wiley Publishing Asia Pty Ltd and Chinese Cochrane Center, West China Hospital of Sichuan University JEBM 6 (2013) 173–184 

175

Impact of pay-for-performance

J. Huang et al.

Table 1 Category of outcome measures QOF indicator ADA domains

Process indicator

Outcome indicator

Glucose control (blood glucose, HbA1c)

Patients with HbA1c record

BP Microvascular events (nephropathy, retinopathy, foot examination)

Patients with BP record Peripheral pulse record; nephropathy testing; microalbuminuria testing; retinal screening; eGFR or sCr testing Patients with total cholesterol record Patients with albuminuria treated with ACEI; prescriptions of ACEI, beta-blocker, aspirin, and statins Patients with flu vaccination

Patients whose HbA1c levels

Impact of pay-for-performance on management of diabetes: a systematic review.

To review and synthesize published evidence of pay-for-performance (P4P) effects on management of diabetes...
346KB Sizes 0 Downloads 0 Views