Olakunle Alonge

Implementation research: what it is and how to do it.

Implementation research is a growing but not well understood field of health research that can contribute to more effective public health and clinical policies and programmes. This article provides a broad definition of implementation research and outlines key principles for how to do it

Republished research: Implementation research: what it is and how to do it: implementation research is a growing but not well understood field of health research that can contribute to more effective public health and clinical policies and programmes. This article provides a broad definition of implementation research and outlines key principles for how to do it.

Implementation research is a growing but not well understood field of health research that can contribute to more effective public health and clinical policies and programmes. This article provides a broad definition of implementation research and outlines key principles for how to do it

Reducing the global burden of childhood unintentional injuries

Among 1–19-year olds, unintentional injuries accounted for 12% of 5.1 million global deaths from injuries in 2010. Despite this high burden, childhood injuries have not received much attention in global health. This paper describes the major causes of deaths from childhood unintentional injuries and provides a review of interventions for reducing this burden. About 627 741 deaths were due to unintentional injuries in 2010 among 1–19-year olds. The proportionate mortality increased with age—from 12.6% among 1–4-year olds to 28.8% among 15–19-year olds. Deaths from Western sub-Saharan Africa and South Asia accounted for more than 50% of all deaths. Rates in these regions are 68.0 and 36.4 per 100 000 population, respectively, compared to 6.4 in Western Europe. Road traffic injuries (RTI) are the commonest cause of death, followed by deaths from drowning, burns and falls. Male children are more predisposed to unintentional injuries except for burns which occur more frequently among females in low and middle income countries (LMICs). Effective solutions exist—including barriers for preventing drowning; safer stoves for burns; child restraint systems for RTI—but the effectiveness of these measures need to be rigorously tested in LMICs. The general lack of a coordinated global response to the burden of childhood unintentional injuries is of concern. The global community must create stronger coalitions and national or local plans for action. Death rates for this paper may have been underestimated, and there is need for longitudinal studies to accurately measure the impact of injuries in LMICs.

Effectiveness of a pay-for-performance intervention to improve maternal and child health services in Afghanistan: a cluster-randomized trial

BACKGROUND A cluster randomized trial of a pay-for-performance (P4P) scheme was implemented in Afghanistan to test whether P4P could improve maternal and child (MCH) services. METHODS All 442 primary care facilities in 11 provinces were matched by type of facility and outpatient volume, and randomly assigned to the P4P or comparison arm. P4P facilities were given bonus payments based on the MCH services provided. An endline household sample survey was conducted in 72 randomly selected matched pair catchment areas (3421 P4P households; 3427 comparison).The quality of services was assessed in 81 randomly sampled matched pairs of facilities. Data collectors and households were blinded to the intervention assignment. MCH outcomes were assessed at the cluster level. RESULTS There were no substantial differences in any of the five MCH coverage indicators (P4P vs comparison): modern contraception(10.7% vs 11.2% (P = 0.90); antenatal care: 56.2% vs 55.6% (P = 0.94); skilled birth attendance (33.9% vs 28.5%, P = 0.17); postnatal care (31.2% vs 30.3%, P = 0.98); and childhood pentavalent3 vaccination (49.6 vs 52.3%, P = 0.41), or in the equity measures. There were substantial increases in the quality of history and physical examinations index (P = 0.01); client counselling index (P = 0.01); and time spent with patients (P = 0.05). Health workers reported limited understanding about the bonuses. CONCLUSIONS The intervention had minimal effect, possibly due to difficulties communicating with health workers and inattention to demand-side factors. P4P interventions need to consider management and community demand issues.

Assessing the pro-poor effect of different contracting schemes for health services on health facilities in rural Afghanistan

BACKGROUND Despite progress in improving health outcomes in Afghanistan by contracting public health services through non-governmental organizations (NGOs), inequity in access persists between the poor and non-poor. This study examined the distributive effect of different contracting types on primary health services provision between the poor and non-poor in rural Afghanistan. METHOD Contracts to NGOs were made to deliver a common set of primary care services in each province, with the funding agencies determining contract terms. The contracting approaches could be classified into three contracting out types (CO-1, CO-2 and CO-3) and a contracting-in (CI) approach based on the contract terms, design and implementation. Exit interviews of patients attending randomly sampled primary health facilities were collected through systematic sampling across 28 provinces at two time points. The outcome, the odds that a client attending a health facility is poor, was modelled using logistic regression with a robust variance estimator, and the effect of contracting was estimated using the difference-in-difference approach combined with stratified analyses. RESULTS The sample covered 5960 interviews from 306 health facilities in 2005 and 2008. The adjusted odds of a poor client attending a health facility over time increased significantly for facilities under CO-1 and CO-2, with odds ratio of 2.82 (1.49, 5.36) P-value 0.001 and 2.00 (1.33, 3.02) P-value 0.001, respectively. The odds ratios for those under CO-3 and CI were not statistically significantly different over time. When compared with the non-contracting facilities, the adjusted ratio of odds ratios of poor status among clients was significantly higher for only those under CO-1, ratio of 2.50 (1.32, 4.74) P-value 0.005. CONCLUSIONS CO-1 arrangement which allows contractors to decide on how funds are allocated within a fixed lump sum with non-negotiable deliverables, and actively managed through an independent government agency, is effective in improving equity of health services provision.

Saving of Children’s Lives from Drowning Project in Bangladesh

Although childhood deaths from communicable diseases have declined during the last decade, the proportion of deaths resulting from injuries has increased steadily during the same period across the world. Drowning is the second leading cause of injury-related deaths among children worldwide. It accounted for approximately 359,000 deaths in 2011, of which 18% occurred in children between the ages of 1 and 4 years. Thirty-seven percent of the drowning deaths in this age range occur in South Asia region alone,3 comprising six countries, including Afghanistan, Bangladesh, Bhutan, India, Nepal, and Pakistan. Language: en

Epidemiology of Burns in Rural Bangladesh: An Update

Each year, approximately 265,000 deaths occur due to burns on a global scale. In Bangladesh, around 173,000 children under 18 sustain a burn injury. Since most epidemiological studies on burn injuries in low and middle-income countries are based on small-scale surveys or hospital records, this study aims to derive burn mortality and morbidity measures and risk factors at a population level in Bangladesh. A household survey was conducted in seven rural sub-districts of Bangladesh in 2013 to assess injury outcomes. Burn injuries were one of the external causes of injury. Epidemiological characteristics and risk factors were described using descriptive as well as univariate and multivariate logistic regression analyses. The overall mortality and morbidity rates were 2 deaths and 528 injuries per 100,000 populations. Females had a higher burn rate. More than 50% of injuries were seen in adults 25 to 64 years of age. Most injuries occurred in the kitchen while preparing food. 88% of all burns occurred due to flame. Children 1 to 4 years of age were four times more likely to sustain burn injuries as compared to infants. Age-targeted interventions, awareness of first aid protocols, and improvement of acute care management would be potential leads to curb death and disability due to burn injuries.

Our Shrinking Globe: Implications for Child Unintentional Injuries.

Unintentional injuries are a leading cause of deaths for children of all ages. Globally, they accounted for 15.4% of 2.6 million deaths recorded among children aged 1 to 14 years in 2013. The 12 highest burden countries in the world by absolute death count and mortality are low- and middle-income countries (LMIC) except for Russia and Equatorial Guinea. These countries accounted for 58% of the 406,442 unintentional injury deaths among 1 to 14 year olds in 2013. Globalization drives inequalities in the distribution of economic gains, risks, and opportunities for preventing child unintentional injuries between high-income countries and LMIC.

Fatal and non-fatal injury outcomes: results from a purposively sampled census of seven rural subdistricts in Bangladesh

BACKGROUND 90% of the global burden of injuries is borne by low-income and middle-income countries (LMICs). However, details of the injury burden in LMICs are less clear because of the scarcity of data and population-based studies. The Saving of Lives from Drowning project, implemented in rural Bangladesh, did a census on 1·2 million people to fill this gap. This Article describes the epidemiology of fatal and non-fatal injuries from the study. METHODS In this study, we used data from the baseline census conducted as part of the Saving of Lives from Drowning (SoLiD) project. The census was implemented in 51 unions from seven purposively sampled rural subdistricts of Bangladesh between June and November, 2013. Sociodemographic, injury mortality, and morbidity information were collected for the whole population in the study area. We analysed the data for descriptive measures of fatal and non-fatal injury outcomes. Age and gender distribution, socioeconomic characteristics, and injury characteristics such as external cause, intent, location, and body part affected were reported for all injury outcomes. FINDINGS The census covered a population of 1 169 593 from 270 387 households and 451 villages. The overall injury mortality rate was 38 deaths per 100 000 population per year, and 104 703 people sustained major non-fatal injuries over a 6-month recall period. Drowning was the leading external cause of injury death for all ages, and falls caused the most number of non-fatal injuries. Fatal injury rates were highest in children aged 1-4 years. Non-fatal injury rates were also highest in children aged 1-4 years and those aged 65 years and older. Males had more fatal and non-fatal injuries than females across all external causes except for burns. Suicide was the leading cause of injury deaths in individuals aged 15-24 years, and more than 50% of the suicides occurred in females. The home environment was the most common location for most injuries. INTERPRETATION The burden of fatal and non-fatal injuries in rural Bangladesh is substantial, accounting for 44 050 deaths and 21 million people suffering major events annually. Targeted approaches addressing drowning in children (especially those aged 1-4 years), falls among the elderly, and suicide among young female adults are urgently needed to reduce injury deaths and morbidity in Bangladesh. FUNDING Bloomberg Philanthropies.

Utility and limitations of measures of health inequities: a theoretical perspective.

Summary box What is already known on this subject? Various measures have been used in quantifying health inequities among populations in recent times; most of these measures were derived to capture the socioeconomic inequalities in health. These different measures do not always lend themselves to common interpretation by policy makers and health managers because they each reflect limited aspects of the concept of health inequities. What does this study add? To inform a more appropriate application of the different measures currently used in quantifying health inequities, this article explicates common theories underlying the definition of health inequities and uses this understanding to show the utility and limitations of these different measures. It also suggests some key features of an ideal indicator based on the conceptual understanding, with the hope of influencing future efforts in developing more robust measures of health inequities. The article also provides a conceptual ‘product label’ for the common measures of health inequities to guide users and ‘consumers’ in making more robust inferences and conclusions. This paper examines common approaches for quantifying health inequities and assesses the extent to which they incorporate key theories necessary for explicating the definition of health inequity. The first theoretical analysis examined the distinction between inter-individual and inter-group health inequalities as measures of health inequities. The second analysis considered the notion of fairness in health inequalities from different philosophical perspectives. To understand the extent to which different measures of health inequities incorporate these theoretical explanations, four criteria were used to assess each measure: 1) Does the indicator demonstrate inter-group or inter-individual health inequalities or both; 2) Does it reflect health inequalities in relation to socioeconomic position; 3) Is it sensitive to the absolute transfer of health (outcomes, services, or both) or income/wealth between groups; 4) Could it be used to capture inequalities in relation to other population groupings (other than socioeconomic status)? The measures assessed include: before and after measures within only the disadvantaged population, range, Gini coefficient, Pseudo-Gini coefficient, index of dissimilarity, concentration index, slope and relative indices of inequality, and regression techniques. None of these measures satisfied all the four criteria, except the range. Whereas each measure quantifies a different perspective in health inequities, using a measure within only the disadvantaged population does not measure health inequities in a meaningful way, even using before and after changes. For a more complete assessment of how programs affect health inequities, it may be useful to use more than one measure.