Who might support or oppose our proposal?
Phelps Charles E, Parente Stephen T in The Economics of US Health Care Policy, 2017
A Quasi-Experiment in Massachusetts. A more recent analysis pertinent to this issue reached a different conclusion. This 2014 “quasi-experimental” study (as self-described by the authors) analyzed the effects on population mortality of the expansion in insurance coverage arising from the Massachusetts “Health Care for All” legislation.19 They looked at county-level death rates in Massachusetts for 2001–2005 (before the health care law came into being) and then 2007–2010. They compared the differences in mortality (county by county) with similar data from carefully matched “control” counties. They found a significant decrease in all-cause mortality associated with the Massachusetts reform, estimated at a 2.9% reduction in all-cause deaths (8.9 deaths per 100,000 adults per year), and a 4.5% reduction in health care-amenable deaths. Their approach is known in the economic literature as a “difference of differences” approach, designed to eliminate effects of other things that might have occurred in the before/after time comparison.
Nonlinear Latent Growth Curve Models
Jason T. Newsom in Longitudinal Structural Equation Modeling, 2015
The loadings for the differential structural equation model are obtained from a simple method of approximating the first and second derivative, called local linear approximation. The essential idea can be stated in terms of differences. A linear slope can be described in terms of the first derivative, where the slope is a ratio of the change in y between any two points relative to the change in some time metric, t, The concept of the derivative, however, is to estimate this ratio for an infinitesimally small increment, so theoretically Δ represents a difference between two points on a curve that approaches, but does not reach, a limit of 0. As we know, the linear slope gives information about the rate of change. A simplified description of the second derivative can be stated in terms of two sets of differences. For example, if the first two points mentioned are said to be Δy2−1 and Δt2−1, then, using the same type of difference between the subsequent adjacent points, the local linear approximation of the second derivative is . This “difference between differences” captures the change in the rate over time or acceleration. (Appendix B provides a brief introduction to derivatives.)
Alternative Study Designs
Richard J. Hayes, Lawrence H. Moulton in Cluster Randomised Trials, 2017
where is the prevalence at time-point t in treatment arm i. This interaction effect contains four random quantities, with correspondingly much larger variance than a final status comparison (see Section 8.2.2.2). If measurement of changes in intervention effect is an important objective of the trial, sample sizes need to be chosen specifically to address this objective. Note that this is conceptually different from using baseline data in a difference-in-differences analysis to reduce variability in estimation of a treatment effect (see Section 7.6.6).
Does Non-Targeted Community CPR Training Increase Bystander CPR Frequency?
Published in Prehospital Emergency Care, 2018
Amy Uber, Richard C. Sadler, Todd Chassee, Joshua C. Reynolds
Continuous variables were summarized according to distribution and were compared between the 2 periods using the t-test. Categorical variables were summarized by the frequency and proportion and were compared with the chi-square or Fisher exact test. We also tested for differences in clinical variables and outcomes stratified by location of OHCA within or outside of a Thiessen polygon. Difference-in-differences analysis (17) with GEE modeling assessed the proportion of OHCA cases with bystander CPR before and after the intervention, adjusting for covariates (age, sex, witnessed, shockable rhythm, public location), and interactions with location within/outside a Thiessen polygon. Difference-in-differences analyses is often used in longitudinal observational studies in order to assess the impact of an intervention while accounting for any background differences that may occur during the study period. Similar modeling tested for changes in hand-only CPR, ROSC, survival to hospital discharge, and favorable neurologic status (CPC 1–2) at hospital discharge. Statistical analyses were performed with STATA 12.0 (StataCorp, College Station, TX) utilizing a 2-sided alpha of 0.05.
Primary care practice-based care management for chronically ill patients (PraCMan) in German healthcare: Outcome of a propensity-score matched cohort study
Published in European Journal of General Practice, 2021
Jonas D. Senft, Tobias Freund, Michel Wensing, Simon Schwill, Regina Poss-Doering, Joachim Szecsenyi, Gunter Laux
In the one-year period before intervention began the per-patient hospitalisation rate was 3.5% higher for patients in the intervention group compared to control (p < 0.05). In the one-year period 12 months after beginning of the intervention the per-patient hospitalisation rate was 8.3% lower in the intervention group than control (p < 0.001). Per-patient hospitalisation costs in the pre-interventional period were 3.1% higher for patients in the intervention group compared to control (p: n.s.). In the observation period after beginning of the intervention per-patient hospitalisation costs were 9.4% lower in favour of the intervention group (p < 0.001). Average per-patient hospitalisation rates and costs for the observation periods are shown in Tables 2 and 3. The ‘Difference in Differences’ shows the difference between cases and controls regarding their particular change from the year before intervention to the year after intervention. In 2017, we had an overall average hospitalisation rate of 0.287 and average hospitalisation costs of 1243€per patient participating in the specific primary care-centred programme (‘Hausarztzentrierte Versorgung,’ HZV) for N = 1,037,093 patients. No significant difference was found with regard to the mortality rate at the end of the one-year observation period (cases: 11.87%, controls: 11.88%, p: n.s.).
Effects of the COVID-19 pandemic on routine pediatric vaccination in Brazil
Published in Expert Review of Vaccines, 2021
Victor Santana Santos, Sarah Cristina Fontes Vieira, Ikaro Daniel de Carvalho Barreto, Vanessa Tavares de Gois-Santos, Ariel Oliveira Celestino, Carla Domingues, Luis Eduardo Cuevas, Ricardo Queiroz Gurgel
The percentage differences of vaccine doses administered during the pre-pandemic, stay-at-home and reopening periods were calculated by comparing them to the same months in 2019. To control for period changes, we conducted difference-in-difference (DiD) analyses and estimated adjusted percentage differences and 95% confidence intervals (CIs) during the stay-at-home and reopening periods using Poisson regression models, adjusting for the percentage difference in vaccine doses administered during the pre-pandemic period in 2020. The difference-in-differences approach uses repeated cross-sectional data collected before and after an event. The difference-in-differences estimator corresponds to the difference in two before-after differences observed in an exposed and non-exposed group. In our analyses, the exposed condition was represented by the months from January to December of the pandemic year 2020, while January to December 2019 were our non-exposed condition. We used the software R Core Team 2021 (Version 4.0.4) and adopted 5% significant level in all analysis.
Related Knowledge Centers
- Observational Study
- Selection Bias
- Time Series
- Experiment
- Treatment & Control Groups
- Natural Experiment
- Dependent & Independent Variables
- Omitted-Variable Bias
- Sampling
- Matching