Assessing the impact of natural policy experiments on socioeconomic inequalities in health: how to apply commonly used quantitative analytical methods?

Long-Term Effects of Integrated Strategies of Community Health Promotion on Diabetes Mellitus Mortality: a Natural Policy Experiment Based on Aggregated Longitudinal Secondary Data

Integrated strategies of community health promotion (ISCHP) are based on intersectoral collaborations using the Health in All Policies approach to address determinants of health. While effects on health determinants have been shown, evidence on the effectiveness of ISCHP on health outcomes is scarce. The aim of this study is to assess the long-term effects of ISCHP on diabetes mellitus mortality (DMM) in German communities. A nonrandomized evaluation based on secondary county-level official data (1998-2016) was performed. In April 2019, 149 communities in Germany with ISCHP out of 401 were identified. Communities with < 5 measurements of DMM, starting before 1999 or after 2015, were excluded. Analyses included 65 communities with ISCHP (IG) and 124 without ISCHP (CG). ISCHP ran for a mean of 5.6 years. Fixed effects (FE) models were used to estimate effects of ISCHP and duration on DMM taking into account the time-varying average age. The FE estimator for DMM is b =  - 2.48 (95% CI - 3.45 to - 1.51) for IG vs. CG and b =  - 0.30 (95% CI - 0.46 to - 0.14) for ISCHP duration (0-16 years). In the first year of an ISCHP, a reduction of the annual DMM of 0.3 per 100,000 population (1%), and in the 16th year of 4.8 (14%) was achieved. This study provides preliminary evidence of the effectiveness of ISCHP in Germany. Limitations include inaccuracies to classify IG and CG and possible selection bias. Longitudinal county-level data may be an efficient data source to evaluate complex interventions, thereby contributing to further strengthen evidence-based integrated health promotion.

Effects of social network diversity in the disablement process: a comparison of causal inference methods and an outcome-wide approach to the Indonesian Family Life Surveys, 2007-2015

BACKGROUND

Social networks (SN) have been proven to be instrumental for healthy aging and function as important safety nets, particular for older adults in low and middle-income countries (LMICs). Despite the importance of interpreting health outcomes in terms of SN, in many LMICs - including Indonesia - epidemiological studies and policy responses on the health effects of SN for aging populations are still uncommon. Using outcome-wide multi-method approaches to longitudinal panel data, this study aims to outline more clearly the role of SN diversity in the aging process in Indonesia. We explore whether and to what degree there is an association of SN diversity with adult health outcomes and investigate potential gender differences, heterogeneous treatment effects, and effect gradients along disablement processes.

METHODS

Data came from the fourth and fifth waves of the Indonesian Family Life Survey fielded in 2007-08 and 2014-15. The analytic sample consisted of 3060 adults aged 50+ years. The primary exposure variable was the diversity of respondents' SN at baseline. This was measured through a social network index (SNI), conjoining information about household size together with a range of social ties with whom respondents had active contact across six different types of role relationships. Guided by the disablement process model, a battery of 19 outcomes (8 pathologies, 5 impairments, 4 functional limitations, 2 disabilities) were included into analyses. Evidence for causal effects of SN diversity on health was evaluated using outcome-wide multivariable regression adjustment (RA), propensity score matching (PSM), and instrumental variable (IV) analyses.

RESULTS

At baseline, 60% of respondents had a low SNI. Results from the RA and PSM models showed greatest concordance and that among women a diverse SN was positively associated with pulmonary outcomes and upper and lower body functions. Both men and women with a high SNI reported less limitations in performing activities of daily living (ADL) and instrumental ADL (IADL) tasks. A high SNI was negatively associated with C-reactive protein levels in women. The IV analyses yielded positive associations with cognitive functions for both men and women.

CONCLUSIONS

Diverse SN confer a wide range of strong and heterogeneous long-term health effects, particularly for older women. In settings with limited formal welfare protection, intervening in the SN of older adults and safeguarding their access to diverse networks can be an investment in population health, with manifold implications for health and public policy.

Sustainable, healthy cities: protocol of a mixed methods evaluation of a cluster randomized controlled trial for Aedes control in Brazil using a community mobilization approach

Background

Dengue is increasing in its global presence with an estimated 4 billion people at-risk of infection in at least 128 countries. Despite the promising results of EcoHealth and community mobilization approaches to Aedes reduction, more evidence of their efficacy on reducing dengue risk is needed. The principal research question is to determine if interventions based upon community mobilization reduce the risk of dengue virus infection among children 3 to 9 years old compared to usual dengue control practice in Fortaleza, Brazil.

Methods

The present study will follow a pragmatic cluster randomized controlled trial (cRCT) design with randomization at the census tract level with equal allocation to the two arms. In each arm, there will be 34 clusters of 86 children between 3 to 9 years old for an expected total of 5848 children enrolled in the study, assuming a risk reduction of 29.5% based upon findings from a previous multi-site cRCT. The primary outcomes are rates of anti-dengue Immunoglobulin G (IgG) seroconversion and adult female Aedes density. The intervention is based upon a participatory health research approach, Socializing Evidence for Participatory Action (SEPA), where the research evidence is used to foster community engagement and ownership of the health issue and solution. Following allocation, intervention communities will develop and implement their own solutions that will likely include a wide variety of collective events and media approaches. Data collection activities over a period of 3 years include household visits for blood collection, household surveys, and entomological surveys; and qualitative activities including focus groups, in-depth interviews, and document analysis to evaluate the process, acceptability, fidelity, and sustainability of the intervention. Study participants will be aware of their assignment and all research staff will be blinded although the intervention assignment will likely be revealed to field staff through interaction with participants.

Discussion

The results of our study will provide evidence on community mobilization as an intervention for dengue control. We anticipate that if community mobilization is effective in Fortaleza, the results of this study will help develop evidence-based vector control programs in Brazil, and also in other countries struggling with Aedes-transmitted diseases.

Trial registration

ISRCTN66131315, registration date: 1 October 2018.

Did an urban perinatal health programme in Rotterdam, the Netherlands, reduce adverse perinatal outcomes? Register-based retrospective cohort study

OBJECTIVES

To study the effect of an urban perinatal health programme in Rotterdam, the Netherlands, on perinatal outcomes.

DESIGN

A retrospective cohort study with difference-in-differences analysis using individual-level perinatal outcome data from the Dutch Perinatal Registry 2003-2014 linked to Central Bureau of Statistics data of migration background and individual disposable household income.

INTERVENTION

The programme consisted of perinatal health promotion, risk selection and risk-guided pregnancy care, and a new primary care child birth centre. The programme was implemented during 2009-2012.

PRIMARY OUTCOME MEASURES

We compared trends in perinatal mortality, preterm delivery and small-for-gestational-age births between targeted urban neighbourhoods in Rotterdam (n=61 415) and all other urban neighbourhoods in the Netherlands (n=881 202). The effect of the programme was modelled as a change in trend of each perinatal outcome in the treatment group post intervention compared with the control population from January 2010 onwards. All analyses were adjusted for maternal age, parity, ethnicity and individual-level low socioeconomic status (SES). We also conducted a stratified analysis by SES.

RESULTS

During 2003-2014, downward trends in perinatal mortality (adjusted OR (aOR) 0.9439 per year, 95% CI 0.9362 to 0.9517), preterm birth (aOR 0.9970 per year, 95% CI 0.9944 to 0.9997) and small-for-gestational-age births (aOR 0.9809 per year, 95% CI 0.9787 to 0.9831) in the entire study population were observed. No demonstrable changes in these trends were found in the intervention group after the programme had started. The stratified analyses by SES showed no changes in trends post intervention in both strata either.

CONCLUSIONS

The programme had no demonstrable effects on perinatal outcomes. The intervention may not have reached a sufficient proportion of the population or has provided too little contrast to the widespread attention for inequalities in pregnancy outcomes occurring simultaneously in the Netherlands.

Is There a Difference in the Utilisation of Inpatient Services Between Two Typical Payment Methods of Health Insurance? Evidence from the New Rural Cooperative Medical Scheme in China

This study aimed to evaluate the effects of the differences between two typical payment methods for the new rural cooperative medical scheme (NRCMS) in China on the utilisation of inpatient services. Interrupted time-series analysis (ITSA) and propensity score matching (PSM) were used to measure the difference between two typical payment methods for the NRCMS with regard to the utilisation of inpatient services. After the reform was formally implemented, the level and slope difference after reform compared with pre-intervention (distribution of inpatients in county hospitals (DIC), distribution of inpatients in township hospitals (DIT) and the actual compensation ratio of inpatients (ARCI)) were not statistically significant. Kernel matching obtained better results in reducing the mean and median of the absolute standardised bias of covariates of appropriateness of admission (AA), appropriateness of disease (AD). The difference in AA and AD of the matched inpatients between two groups was −0.03 (p-value = 0.042, 95% CI: −0.08 to 0.02) and 0.21 (p-value < 0.001, 95% CI: −0.17 to 0.25), respectively. The differences in the utilisation of inpatient services may arise owing to the system designs of different payment methods for NRCMS in China. The causes of these differences can be used to guide inpatients to better use medical services, through the transformation and integration of payment systems.

Commentary: Causal Inference for Social Exposures

Social epidemiology seeks to describe and quantify the causal effects of social institutions, interactions, and structures on human health. To accomplish this task, we define exposures as treatments and posit populations exposed or unexposed to these well-defined regimens. This inferential structure allows us to unambiguously estimate and interpret quantitative causal parameters and to investigate how these may be affected by biases such as confounding. This paradigm has been challenged recently by some critics who favor broadening the exposures that may be studied beyond treatments to also consider states. Defining the exposure protocol of an observational study is a continuum of specificity, and one may choose to loosen this definition, incurring the cost of causal parameters that become commensurately more vague. The advantages and disadvantages of broader versus narrower definitions of exposure are matters of continuing debate in social epidemiology as in other branches of epidemiology.

Effect of integrated urban and rural residents medical insurance on the utilisation of medical services by residents in China: a propensity score matching with difference-in-differences regression approach

OBJECTIVES

In this study, we aim to evaluate the effect of urban and rural resident medical insurance scheme (URRMI) on the utilisation of medical services by urban and rural residents in the four pilot provinces.

SETTING AND PARTICIPANTS

The sample used in this study is 13 305 individuals, including 2620 in the treatment group and 10 685 in the control group, from the 2011 and 2015 surveys of China Health and Retirement Longitudinal Study.

OUTCOME MEASURES

Propensity score matching and difference-in-differences regression approach (PSM-DID) is used in the study. First, we match the baseline data by using kernel matching. Then, the average treatment effect of the four outcome variables are analysed by using the DID model. Finally, the robustness of the PSM-DID estimation is tested by simple model and radius matching.

RESULTS

Kernel matching have improved the overall balance after matching. The URRMI policy has significantly reduced the need-but-not outpatient care and significantly increased outpatient care cost and inpatient care cost for rural residents, with DID value of -0.271, 0.090 and 0.256, respectively. After robustness test, the DID competing results of four outcome variables are consistent.

CONCLUSIONS

URRMI has a limited effect on the utilisation of medical and health services by all residents, but the effect on rural residents is obvious. The government should establish a unified or income-matching payment standard to prevent, control the use of medical insurance funds and increase its efforts to implement URRMI integration in more regions to improve overall fundraising levels.

Impact of Public Smoking Bans on Social Inequalities in Children’s Exposure to Tobacco Smoke at Home: An Equity-Focused Systematic Review

Introduction

A systematic review was conducted to evaluate the impact of public smoking bans on social inequalities in children’s secondhand smoke (SHS) exposure at home.

Methods

Five databases were electronically searched for articles on children’s SHS exposure at home related to public smoking bans. In addition, the gray literature and German public health journals were considered. Search was restricted to English and German publications. Of 3037 records screened, 25 studies fulfilled the inclusion criteria by either measuring SHS exposure before and after public smoking ban introduction or by comparing exposure between regions with and without smoke-free legislation. Studies were further examined whether they additionally reported on impacts on social inequalities in SHS exposure. Information on children’s SHS exposure at home in relation to smoke-free legislation were extracted by one reviewer and checked for accuracy by a second reviewer. According to Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Equity (PRISMA-E) guidelines for equity-focused systematic reviews, the PROGRESS-Plus framework was applied to data extraction and analysis with focus on social inequalities in SHS exposure. Results were visualized by a harvest plot.

Results

Eight studies gave results on the impact of public smoking bans on social inequalities in children’s SHS exposure. Whereas only one study indicated widening of the social gap in exposure, seven studies showed no impact or a reduction of social inequalities in exposure.

Conclusions

First evidence on short-term impact of public smoking bans does not support the assumption of intervention-generated inequalities in children’s SHS exposure at home. Future studies should focus on long-term equity impacts of smoke-free legislation.

Implications

There are substantial social inequalities in children’s SHS exposure in many countries. Both hypotheses on the effect of smoke-free legislation on children’s SHS exposure at home, the displacement hypothesis and the social diffusion hypothesis, did not take social inequalities into account. Up to now, only few studies analyzed the effects of smoke-free legislation on social inequalities in children’s SHS exposure at home. Public smoking bans had overall no negative impact on social inequalities in children’s SHS exposure at home. More consistent reporting of absolute and relative inequalities is needed to comprehensively assess equity impact of smoke-free legislation.

Addressing Health Equity Through Action on the Social Determinants of Health: A Global Review of Policy Outcome Evaluation Methods

Background: Epidemiological evidence on the social determinants of health inequity is well-advanced, but considerably less attention has been given to evaluating the impact of public policies addressing those social determinants. Methodological challenges to produce evidence on policy outcomes present a significant barrier to mobilising policy actions for health equities. This review aims to examine methodological approaches to policy evaluation of health equity outcomes and identify promising approaches for future research.

Methods: We conducted a systematic narrative review of literature critically evaluating policy impact on health equity, synthesizing information on the methodological approaches used. We searched and screened records from five electronic databases, using pre-defined protocols resulting in a total of 50 studies included for review. We coded the studies according to (1) type of policy analysed; (2) research design; (3) analytical techniques; (4) health outcomes; and (5) equity dimensions evaluated.

Results: We found a growing number of a wide range of policies being evaluated for health equity outcomes using a variety of research designs. The majority of studies employed an observational research design, most of which were cross-sectional, however, other approaches included experimental designs, simulation modelling, and meta-analysis. Regression techniques dominated the analytical approaches, although a number of novel techniques were used which may offer advantages over traditional regression analysis for the study of distributional impacts of policy. Few studies made intra-national or cross-national comparisons or collected primary data. Despite longstanding challenges of attribution in policy outcome evaluation, the majority of the studies attributed change in physical or mental health outcomes to the policy being evaluated.

Conclusion: Our review provides an overview of methodological approaches to health equity policy outcome evaluation, demonstrating what is most commonplace and opportunities from novel approaches. We found the number of studies evaluating the impacts of public policies on health equity are on the rise, but this area of policy evaluation still requires more attention given growing inequities.