Estimating benefits and harms of screening across subgroups: the Canadian Task Force on Preventive Health Care integrates the GRADE approach and overcomes minor challenges

Interventions to optimize embryo transfer in women undergoing assisted conception: a comprehensive systematic review and meta-analyses

BACKGROUND

Several interventions and techniques are suggested to improve the outcome of embryo transfer (ET) in assisted conception. However, there remains no consensus on the optimal practice, with high variations among fertility specialists.

OBJECTIVE AND RATIONALE

We conducted a comprehensive systematic review and meta-analyses of randomized controlled trials (RCTs) aiming to identify effective interventions that could be introduced around the time of ET to improve reproductive outcomes.

SEARCH METHODS

We searched the electronic databases (MEDLINE, EMBASE and Cochrane CENTRAL) from inception until March 2021 using a multi-stage search strategy of MeSH terms and keywords, and included all RCTs that evaluated an intervention in the 24-h period before/after ET in women undergoing IVF/ICSI. Our primary outcome was clinical pregnancy rate post-ET confirmed as viable pregnancy on ultrasound scan. We assessed the risk of bias in included trials and extracted data in duplicate. We pooled data using a random-effect meta-analysis and reported using risk ratio (RR) with 95% CI. We explored publication bias and effect modifiers using subgroup analyses.

OUTCOMES

Our search yielded 3685 citations of which we included 188 RCTs (38 interventions, 59 530 participants) with a median sample size of 200 (range 26-1761). The quality of included RCTs was moderate with most showing a low risk of bias for randomization (118/188, 62.8%) and attrition (105/188, 55.8%) but there was a significant risk of publication bias (Egger's test P = 0.001). Performing ET with ultrasound guidance versus clinical touch (n = 24, RR 1.265, 95% CI 1.151-1.391, I2 = 38.53%), hyaluronic acid versus routine care (n = 9, RR 1.457, 95% CI 1.197-1.261, I2 = 46.48%) and the use of a soft versus hard catheter (n = 27, RR 1.122, 95% CI 1.028-1.224, I2 = 57.66%) led to higher clinical pregnancy rates. Other pharmacological add-ons also showed a beneficial effect including granulocyte colony-stimulating factor (G-CSF: n = 4, RR 1.774, 95% CI 1.252-2.512, I2 = 0), Atosiban (n = 7, RR 1.493, 95% CI 1.184-1.882, I2 = 68.27%) and hCG (n = 17, RR 1.232, 95% CI 1.099-1.382, I2 = 57.76%). Bed rest following ET was associated with a reduction in clinical pregnancy (n = 6, RR 0.857, 95% CI 0.741-0.991, I2 = 0.01%). Other commonly used interventions, such as non-steroidal anti-inflammatory drugs, prophylactic antibiotics, acupuncture and cervical mucus removal, did not show a significant benefit on reproductive outcomes. Our effect estimates for other important outcomes, including miscarriage and live birth, were limited by the varied reporting across included RCTs.

WIDER IMPLICATIONS

Using ultrasound guidance, soft catheters and hyaluronic acid at the time of ET appears to increase clinical pregnancy rates. The use of Atosiban, G-CSF and hCG showed a trend towards increased clinical pregnancy rate, but larger trials are required before adopting these interventions in clinical practice. Bed rest post-ET was associated with a reduction in clinical pregnancy and should not be recommended.

Better guidelines for better care: accounting for multimorbidity in clinical guidelines – structured examination of exemplar guidelines and health economic modelling

Background

Multimorbidity is common but most clinical guidelines focus on single diseases.

Aim

To test the feasibility of new approaches to developing single-disease guidelines to better account for multimorbidity.

Design

Literature-based and economic modelling project focused on areas where multimorbidity makes guideline application problematic.

Methods

(1) Examination of accounting for multimorbidity in three exemplar National Institute for Health and Care Excellence guidelines (type 2 diabetes, depression, heart failure); (2) examination of the applicability of evidence in multimorbidity for the exemplar conditions; (3) exploration of methods for comparing absolute benefit of treatment; (4) incorporation of treatment pay-off time and competing risk of death in an exemplar economic model for long-term preventative treatments with slowly accruing benefit; and (5) development of a discrete event simulation model-based cost-effectiveness analysis for people with both depression and coronary heart disease.

Results

(1) Comorbidity was rarely accounted for in the clinical research questions that framed the development of the exemplar guidelines, and was rarely accounted for in treatment recommendations. Drug–disease interactions were common only for comorbid chronic kidney disease, but potentially serious drug–drug interactions between recommended drugs were common and rarely accounted for in guidelines. (2) For all three conditions, the trials underpinning treatment recommendations largely excluded older, more comorbid and more coprescribed patients. The implications of low applicability varied by condition, with type 2 diabetes having large differences in comorbidity, whereas potentially serious drug–drug interactions were more important for depression. (3) Comparing absolute benefit of treatments for different conditions was shown to be technically feasible, but only if guideline developers are willing to make a number of significant assumptions. (4) The lifetime absolute benefit of statins for primary prevention is highly sensitive to the presence of both the direct treatment disutility of taking a daily tablet and competing risk of death. (5) It was feasible to use a discrete event simulation-based model to represent the relevant care pathways to estimate the relative cost-effectiveness of pharmacological treatments of major depressive disorder in primary care for patients who are also likely to go on and receive treatment for coronary heart disease but the analysis was reliant on eliciting some parameter values from experts, which increases the inherent uncertainty in the results. The key limitation was that real-life use in guideline development was not examined.

Conclusions

Guideline developers could feasibly (1) use epidemiological data characterising the guideline population to inform consideration of applicability and interactions; (2) systematically compare the absolute benefit of long-term preventative treatments to inform decision-making in people with multimorbidity and high treatment burden; and (3) modify the output from economic models used in guideline development to examine time to benefit in terms of the pay-off time and varying competing risk of death from other conditions.

Future work

Further research is needed to optimise presentation of comparative absolute benefit information to clinicians and patients, to evaluate the use of epidemiological and time-to-benefit data in guideline development, to better quantify direct treatment disutility and to better quantify benefit and harm in people with multimorbidity.

Funding

The National Institute for Health Research Health Services and Delivery Research programme.