Using Split Samples and Evidence Factors in an Observational Study of Neonatal Outcomes

Zhang, Kai; Small, Dylan S.; Lorch, Scott A.; Srinivas, Sindhu K.; Rosenbaum, Paul R.

doi:10.1198/jasa.2011.ap10604

Cited by 32 publications

(29 citation statements)

References 48 publications

(52 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[22][23][24] Zhang and colleagues 24 found that the closures were associated with a substantial increase in neonatal birth injuries, with increases most pronounced among mothers living in communities with closures. Allen and colleagues 25 compared time trends of iatrogenic preterm delivery, fetal growth restriction and perinatal mortality in regions of Nova Scotia, Canada, Note: CI = confidence interval, IQR = interquartile range.…”

Section: Discussionmentioning

confidence: 99%

Safety of labour and delivery following closures of obstetric services in small community hospitals

Hutcheon

Riddell

Strumpf

et al. 2016

CMAJ

View full text Add to dashboard Cite

The safety of obstetric services in small, rural communities remains uncertain. Delivery at hospitals with low delivery volumes has been correlated with better, worse and comparable pregnancy outcomes compared with delivery at larger centres, [1][2][3][4][5][6][7] and studies examining the safety of delivery at rural versus urban hospitals have likewise produced conflicting results. [8][9][10][11][12] Establishing the relative safety of obstetric care in small rural hospitals is challenging. Studies comparing health outcomes according to hospital delivery volume 3-5 are influenced by underlying referral patterns in which higher rates of adverse outcomes at larger centres may reflect referrals of high-risk women. Adjustment for maternal risk factors is unlikely to completely account for case-mix differences. Comparisons made according to catchment area (in which women are classified according to their place of residence) 1,2,4,[8][9][10][11][12] prevent bias due to referral patterns, but confounding by differences in socioeconomic status and health behaviours between women living in urban and rural areas remains a concern.10 Such comparisons also have less utility for decisions about service regionalization because the policy option is not to change rural women's place of residence to an urban setting, but rather to have rural women travel to an urban setting to deliver, which may lead to increased risks, such as unintentional out-of-hospital delivery. Since 1998, nearly one-third of hospitals in British Columbia, Canada, have stopped providing planned obstetric services. The vast majority of service closures occurred in low-volume hospitals (< 150 deliveries/yr) serving smaller, rural communities. In this study, we examined whether the frequency and severity of maternal-neonatal labour and delivery health outcomes of women residing in small communities were affected by the closure of their community hospitals. Other implications of obstetric service provision in small communities (e.g., social, economic), although important, are beyond the scope of this study. In recent decades, many smaller hospitals in British Columbia, Canada, have stopped providing planned obstetric services. We examined the effect of these service closures on the labour and delivery outcomes of pregnant women living in affected communities.

show abstract

Section: Discussionmentioning

confidence: 99%

Safety of labour and delivery following closures of obstetric services in small community hospitals

Hutcheon

Riddell

Strumpf

et al. 2016

CMAJ

View full text Add to dashboard Cite

show abstract

“…Moreover, these analyses may incorporate sensitivity analyses and may be combined with each other. For detailed discussion of the logic of evidence factors, see Rosenbaum (2010cRosenbaum ( , 2011b and Zhang et al (2011). …”

Section: Practical Advicementioning

confidence: 98%

“…Alternatively, an observational study may be based on administrative records, and then I might be tens of thousands or millions. For instance, Silber et al (2007) used the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) data system in a study with I = 344 matched pairs of patients with ovarian cancer, Zhang et al (2011) studied more than I = 130,000 pairs of mothers using birth records, and Volpp et al (2007) used Medicare claims to study 8.5 million patients in connection with the ACGME resident hours reform. Both design sensitivity and efficiency matter when I = 100, but design sensitivity is decisive as I → ∞.…”

Section: Practical Advicementioning

confidence: 99%

“…Although not evaluated for the choice of a test statistic, Heller, Rosenbaum, and Small (2009) found that sample splitting for I = 1000 matched pairs often improved design choices by more than enough to offset the loss of I/10 = 100 matched pairs; that is, better decisions in design together with 10% less data yielded higher power in a sensitivity analysis. Zhang et al (2011) used a planning sample of about I/10 = 13,000 pairs. As discussed by Cox (1975), splitting offers a benefit that is difficult to quantify, namely the exercise of unautomated human judgment in light of data combined with honest inferences.…”

Section: Practical Advicementioning

confidence: 99%

See 1 more Smart Citation

WeightedM-statistics With Superior Design Sensitivity in Matched Observational Studies With Multiple Controls

Rosenbaum¹

2014

Journal of the American Statistical Association

Self Cite

View full text Add to dashboard Cite

In a nonrandomized or observational study, a weak association between receipt of the treatment and an outcome may be explained not as effects caused by the treatment but rather by a small bias in the assignment of individuals to treatment or control; however, a strong association may be explained as noncausal only by a large bias. The strength of the association between treatment and outcome is not uniform across the data from a study, and this motivates giving greater weight where the association is stronger. In an observational study with treated-control matched pairs, it is known that results are less sensitive to unmeasured biases if pairs with small absolute differences in outcomes are given little weight in the analysis; more precisely, such a test statistic has superior design sensitivity. How should outcomes be weighted if an observational study is matched in sets with one treated subject and several controls? An M-statistic is the quantity equated to zero in defining Huber's M-estimates, including the mean, and it is used in testing hypotheses and setting confidence limits. In matched sets, a weighted M-statistic increases the weight of some matched sets and decreases the weight of others. Not unlike the case of matched pairs, weighted M-statistics with suitable weights have larger design sensitivities, and hence greater power in a sensitivity analysis, than unweighted statistics for symmetric unimodal errors, such as Normal, logistic, or t-distributed errors. This issue is examined using an asymptotic measure, the design sensitivity, and using simulation. For one Normal sampling situation, weighting the matched sets increased the power of a 0.05 level sensitivity analysis from 0.05 without weights to 0.75 with weights. An example from NHANES 2009-2010 concerning methylmercury in the blood of people who consume large amounts of fish is used to illustrate.KEY WORDS: Observational study; Power of a sensitivity analysis.

show abstract

“…Moreover, each of these two tests is only slightly affected by unmeasured biases that might strongly affect the other. That is, the design produces two evidence factors (Rosenbaum 2010a, 2011a; Zhang et al 2011). An optimal matching algorithm yields two nonoverlapping sets of matched pairs.…”

Section: Introduction: Regional or General Anesthesia; Outlinementioning

confidence: 99%

Contrasting Evidence Within and Between Institutions That Provide Treatment in an Observational Study of Alternate Forms of Anesthesia

Zubizarreta

Neuman

Silber

et al. 2012

Journal of the American Statistical Association

Self Cite

View full text Add to dashboard Cite

In a randomized trial, subjects are assigned to treatment or control by the flip of a fair coin. In many nonrandomized or observational studies, subjects find their way to treatment or control in two steps, either or both of which may lead to biased comparisons. By a vague process perhaps affected by proximity or sociodemographic issues, subjects find their way to institutions that provide treatment. Once at such an institution, a second process, perhaps thoughtful and deliberate, assigns individuals to treatment or control. In the current paper, the institutions are hospitals, and the treatment under study is the use of general anesthesia alone versus some use of regional anesthesia during surgery. For a specific operation, the use of regional anesthesia may be typical in one hospital and atypical in another. A new matched design is proposed for studies of this sort, one that creates two types of nonoverlapping matched pairs. Using a new extension of optimal matching with fine balance, pairs of the first type exactly balance treatment assignment across institutions, so each institution appears in the treated group with the same frequency that it appears in the control group; hence, differences between institutions that affect everyone in the same way cannot bias this comparison. Pairs of the second type compare institutions that assign most subjects to treatment and other institutions that assign most subjects to control, so each institution is represented in the treated group if it typically assigns subjects to treatment or alternatively in the control group if it typically assigns subjects to control, and no institution appears in both groups. By and large, in the second type of matched pair, subjects became treated subjects or controls by choosing an institution, not by a thoughtful and deliberate process of selecting subjects for treatment within institutions. The design provides two evidence factors, that is, two tests of the null hypothesis of no treatment effect that are independent when the null hypothesis is true, where each factor is largely unaffected by certain unmeasured biases that could readily invalidate the other factor. The two factors permit separate and combined sensitivity analyses, where the magnitude of bias affecting the two factors may differ. The case of knee surgery in the study of regional versus general anesthesia is considered in detail.

show abstract

Using Split Samples and Evidence Factors in an Observational Study of Neonatal Outcomes

Cited by 32 publications

References 48 publications

Safety of labour and delivery following closures of obstetric services in small community hospitals

Safety of labour and delivery following closures of obstetric services in small community hospitals

WeightedM-statistics With Superior Design Sensitivity in Matched Observational Studies With Multiple Controls

Contrasting Evidence Within and Between Institutions That Provide Treatment in an Observational Study of Alternate Forms of Anesthesia

Contact Info

Product

Resources

About