Interaction and intervention modeling: predicting and extrapolating the impact of multiple interventions

PURPOSE: Methods called interaction and intervention modeling are presented. Interaction modeling examines the interactions between variables as the basis for predicting the impact of multiple variables on a target population and on populations with difference distributions of risk factors. Intervention modeling incorporates these interactions and aims to extrapolate the impact of multiple interventions to new populations. The aim is to develop methods that will be useful for modeling and comparing intervention strategies using existing data and standard statistical methods. METHODS: Traditional hypothesis testing methods used for randomized clinical trials and cohort studies and extrapolating the results to new populations are compared with interaction and intervention modeling methods. Interaction and intervention modeling utilizes the same data as the traditional approach but examines the impact of multiple simultaneous interactions and allows extrapolation of the results to populations with different prevalences and distributions of risk factors. An example using real data demonstrates the potential of interaction and intervention modeling to predict the impact of multiple interacting variables and to compare the impact of alternative interventions. RESULTS: The methods outlined take into account the impact of the magnitude of the relative risks, prevalence of risk factors, and interaction of risk variables when predicting the impact on a new population or extrapolating the results of one or more interventions on a new population. Traditional methods that do not take into account interactions are shown to produce different conclusions from the intervention modeling approach that incorporates interactions. The impact of the intervention modeling approach compared with the traditional approach will be quite variable depending on the prevalence of the risk factors and their extent of interaction. CONCLUSIONS: Studies designed to test a hypothesis treat most variables as potential confounding variables adjusting for their impact and their interactions as part of the analysis using traditional regression methods. Interaction and intervention modeling focuses on the interactions themselves and allows comparison of the effectiveness of alternative interventions.     

Retention of low income mothers in a parenting intervention study

Women with inadequate prenatal care were recruited to a multi-component parenting intervention study. Because it was anticipated that this high-risk population might present challenges to retention, a variety of strategies were employed to maintain their participation in the study. This report reviews the results of these retention efforts and compares the population that completed the study versus those that terminated prior to study completion. Two hundred and eighty-six women were randomized to an intervention or control group. Careful tracking of the mothers, offering incentives for completing various study activities and providing a culturally competent staff were among the strategies employed to maintain participation. Comparison was made of those mothers terminating before study completion versus those retained, and of those terminating early in the study period versus later. Despite retention efforts, attrition at a level of 41% occurred. A few characteristics of mothers terminating early from the study were significant including older maternal age, a larger number of children, and incidence of no prenatal care. Despite comprehensive tracking procedures, some mothers were lost to follow up after change of residence. Other reasons for attrition included child outplacement and refusal of services or data collection procedures.     

Smoking Cessation and Relapse Among Pregnant African-American Smokers in Washington,DC

Smoking is the single most preventable cause of perinatal morbidity. This study examines smoking behaviors during pregnancy in a high risk population of African Americans. The study also examines risk factors associated with smoking behaviors and cessation in response to a cognitive behavioral therapy (CBT) intervention. This study is a secondary analysis of data from a randomized controlled trial addressing multiple risks during pregnancy. Five hundred African-American Washington, DC residents who reported smoking in the 6 months preceding pregnancy were randomized to a CBT intervention. Psycho-social and behavioral data were collected. Self-reported smoking and salivary cotinine levels were measured prenatally and postpartum to assess changes in smoking behavior. Comparisons were made between active smokers and those abstaining at baseline and follow-up in pregnancy and postpartum. Sixty percent of participants reported quitting spontaneously during pregnancy. In regression models, smoking at baseline was associated with older age, <a high school education and illicit drug use. At follow-up closest to delivery, smoking was associated with lower education, smoking and cotinine level at baseline and depression. At postpartum, there was a relapse of 34%. Smokers postpartum were significantly more likely to smoke at baseline and use illicit drugs in pregnancy. Mothers in the CBT intervention were less likely to relapse. African-American women had a high spontaneous quit rate and no response to a CBT intervention during pregnancy. Postpartum mothers’ resolve to maintain a quit status seems to wane despite their prolonged period of cessation. CBT reduced postpartum relapse rates.     

5-NUCLEOTIDASE ACTIVITY OF MEGAKARYOBLASTS IN A CASE OF ACUTE MEGAKARYOBLASTIC LEUKAEMIA

S ummary . The enzyme 5'-nucleotidase (5'-N) was demonstrated cytochemi-cally in the blast cells from a case of acute megakaryoblastic leukaemia. No activity of 5'-N could be detected in blasts of other types of acute leukaemia.     

Very Preterm Birth is Reduced in Women Receiving an Integrated Behavioral Intervention: A Randomized Controlled Trial

This study examines whether an integrated behavioral intervention with proven efficacy in reducing psycho-behavioral risks (smoking, environmental tobacco smoke exposure (ETSE), depression, and intimate partner violence (IPV)) in African-Americans is associated with improved pregnancy outcomes. A randomized controlled trial targeting risks during pregnancy was conducted in the District of Columbia. African-American women were recruited if reporting at least one of the risks mentioned above. Randomization to intervention or usual care was site and risk specific. Sociodemographic, health risk and pregnancy outcome data were collected. Data on 819 women, and their singleton live born infants were analyzed using an intent-to-treat approach. Bivariate analyses preceded a reduced logistical model approach to elucidate the effect of the intervention on the reduction of prematurity and low birth weight. The incidence of low birthweight (LBW) was 12% and very low birthweight (VLBW) was 1.6%. Multivariate logistic regression results showed that depression was associated with LBW (OR = 1.71, 95% CI = 1.12–2.62). IPV was associated with preterm birth (PTB) and very preterm birth (VPTB) (OR 1.64, 95% CI = 1.07–2.51, OR = 2.94, 95% CI = 1.40–6.16, respectively). The occurrence of VPTB was significantly reduced in the intervention compared to the usual care group (OR = 0.42, 95% CI = 0.19-0.93). Our study confirms the significant associations between multiple psycho-behavioral risks and poor pregnancy outcomes, including LBW and PTB. Our behavioral intervention with demonstrated efficacy in addressing multiple risk factors simultaneously reduced VPTB within an urban minority population.     

An intervention to improve postpartum outcomes in African-American mothers: a randomized controlled trial

OBJECTIVE: To evaluate the efficacy of an integrated multiple risk intervention, delivered mainly during pregnancy, in reducing such risks (cigarette smoking, environmental tobacco smoke exposure, depression, and intimate partner violence) postpartum. METHODS: Data from this randomized controlled trial were collected prenatally and on average 10 weeks postpartum in six prenatal care sites in the District of Columbia. African Americans were screened, recruited, and randomly assigned to the behavioral intervention or usual care. Clinic-based, individually tailored counseling was delivered to intervention women. The outcome measures were number of risks reported postpartum and reduction of these risks between baseline and postpartum. RESULTS: The intervention was effective in significantly reducing the number of risks reported in the postpartum period. In bivariate analyses, the intervention group was more successful in resolving all risks (47% compared with 35%, P=.007, number needed to treat=9, 95% confidence interval [CI] 5-31) and in resolving some risks (63% compared with 54%, P=.009, number needed to treat=11, 95% CI 7-43) as compared with the usual care group. In logistic regression analyses, women in the intervention group were more likely to resolve all risks (odds ratio 1.86, 95% CI 1.25-2.75, number needed to treat=7, 95% CI 4-19) and resolve at least one risk (odds ratio 1.60, 95% CI 1.15-2.22, number needed to treat=9, 95% CI 6-29). CONCLUSION: An integrated multiple risk factor intervention addressing psychosocial and behavioral risks delivered mainly during pregnancy can have beneficial effects in risk reduction postpartum.