You are here
Smoking Prevention Studies
April 23-24, 1998 • Ramada Inn • Bethesda, Maryland
Meta-Analysis of Validated Cessation in Randomized Controlled Trials
Patricia Dolan Mullen, Dr.P.H.
From the Center for Health Promotion Research and Development, University of Texas-Houston School of Public Health, Houston, Texas
Of more than 30 reports of findings from evaluations of pregnancy smoking cessation interventions (PSC) in "developed" countries, 16 were randomized controlled trials (RCTs) with cessation validated after the fifth month. Study populations from this subset of trials included women of high and low socioeconomic status (SES) who were smokers at the time of their first visit. There was little volunteer bias although some smokers undoubtedly chose not to disclose their status.
Most interventions consisted of one or more brief counseling sessions and self-help material tailored for pregnant women. Individuals with backgrounds ranging from public clinic peers to resident physicians provided the counseling. If the counseling offered in the RCTs was classified according to the definitions of counseling intensity used by the AHCPR Smoking Cessation Guideline Panel, two of the studies would be deemed "minimal," that is, counseling sessions lasted for less than 4 minutes, and most of the remainder would be defined as "brief."
At end of pregnancy, cessation rates for the intervention group ranged from 4% to 32%. Typically, the control group received minimal treatment. Their end-of-pregnancy cessation rates ranged from 0% to 17%, with lower percentages found in lower SES populations.
The average effect of these interventions on cessation can be expressed as a risk ratio (RR) or percentage of women stopping smoking in the treatment group divided by the percentage in the control group. The summary RR and the confidence interval (CI) for the 16 studies are 1.7 (CI=1.3-2.2), a 70% improvement in cessation. Subgroup analyses indicate that African-American and Hispanic-American women were no less likely to benefit than white women. Other positive effects included significant, validated reduction and cost savings within 12 months of the intervention investment. The combination of positive effects and low intervention costs provides a strong argument for broad-scale dissemination of brief pregnancy smoking cessation interventions.
It is important to note that these study findings are not homogeneous, and the number of studies is not sufficient to explore sources of variation statistically or qualitatively. In my opinion, there is no compelling case for more testing of the relative effectiveness of these interventions. It appears that we can rule out media alone, decide that patient peers probably are not optimal counselors, and proceed with studies of strategies for diffusion and high-quality implementation by usual care providers–understanding that systems development, including prompts, training, and reinforcement, will be necessary components.
Some of the gaps in our knowledge that can be gleaned from these trials are as follows: (1) disclosure of smoking by women entering prenatal care may be an increasing problem, and this may indicate that use of dipsticks to detect urinary cotinine should be studied ("the new vital sign" in prenatal care?); (2) brief interventions are reaching lighter smokers and not heavier, more addicted smokers, and it is time for serious discussion of the use of pharmacologic therapies during pregnancy; (3) interventions tested to date have not shown an effect in preventing relapse during pregnancy of women who stop smoking spontaneously; (4) fewer low SES women benefit from brief intervention, and interventions to address such issues as stress, family conflict, and depression appear to be worth further exploration; (5) postpartum return to smoking is far too common, and interventions to sustain abstinence achieved during pregnancy should have high priority; (6) cessation interventions directed to fathers who smoke have received too little attention, especially in view of smoking partners’ influence on women’s smoking.
Recommendations for future research include:
- Broad-scale dissemination studies and activities for brief interventions;
- Study of dipsticks to improve case-finding and monitor progress;
- Testing of nicotine replacement therapies for heavier, more addicted smokers;
- Development of interventions to prevent spontaneous quitters from relapsing during pregnancy;
- Testing of interventions to address such issues as stress, family conflict, and depression that may decrease motivation and emotional resources for cessation;
- More development and testing of interventions to sustain abstinence achieved during pregnancy; and
- Development and testing of cessation interventions directed to fathers who smoke.
References for Dr. Mullen’s abstract appear in Appendix B .
Smoking Cessation and Reduction in Pregnancy Project: SCRIPP
Lesa L. Woodby, Ph.D., M.P.H., and Richard A. Windsor, Ph.D., M.P.H.
From the University of Alabama at Birmingham, Birmingham, Alabama
The Smoking Cessation and Reduction in Pregnancy Program (SCRIPP) has been conducted from 1981 to the present in multiple prenatal care clinics involving several thousand patients. Two randomized clinical trials (1981 through 1991), conducted in Alabama in conjunction with the Jefferson County Health Department, evaluated the efficacy of a standardized program of cessation skills and risk counseling appropriate for use in routine prenatal care among high-risk populations. Currently, a third randomized clinical trial (1991-2001) is under way to evaluate the effectiveness of a smoking cessation intervention for pregnant smokers delivered as part of routine care by a representative sample of public health maternity staff across Alabama.
The multicomponent patient education program is based on social cognitive theory. It was developed in consultation with patients, MDs, RNs, MSWs, and RDs. The methods are culturally sensitive and tailored to the pregnant smoker. During the first routine maternity visit, a trained patient educator introduces standardized cessation methods and provides risk counseling. The patient views the video, Commit to Quit, which demonstrates the behavioral modification steps described in A Pregnant Woman’s Guide to Quit Smoking. This self-help guide is given to the woman to assist her in the quitting process. Systematic reinforcement, triggered by a chart reminder, is provided by clinic staff at subsequent routine visits.
Behavioral impact and cost-effectiveness analyses have been performed. SCRIPP increased normal cotinine-confirmed patient quit rates from 2% to 20% and significant reduction rates from 7% to 30%. The estimated cost benefit of dissemination to the annual Alabama Medicaid obstetrical cohort and the U.S. birth cohort and associated clinical impact and excess health care costs for the 1990-2000 decade have been evaluated. Sensitivity analyses indicate that for every $1 spent on SCRIPP smoking cessation methods for pregnant women, an estimated $7 to $17 in medical costs may be saved. This cost reduction does not include the long-term direct and indirect emotional impact experienced by the family.
SCRIPP has also been tested and evaluated in North Carolina, Florida, Michigan, Massachusetts, and Washington, DC. It has been adapted for routine use in Norway, Sweden, Australia, and multiple Canadian provinces. In 1996, a Robert Wood Johnson Foundation dissemination expert panel identified SCRIPP as one of five national programs recommended for routine use in prenatal care as a "best clinical practice method for pregnant smokers." SCRIPP is one of the most efficacious and cost-effective methods available to achieve the Healthy People 2000 managed care health and tobacco prenatal goals for pregnant smokers.
Workgroup members viewed prevention of risk drinking and risk smoking among pregnant women as similar issues and discussed the possibility of applying techniques from the smoking cessation field to the alcohol field, particularly in the areas of nondisclosure and postpartum relapse. According to Dr. Mullen, about 65% of women who quit smoking during pregnancy began again postpartum. Another 25% quit during the first trimester but could not sustain the behavior during the remainder of their pregnancies. To address this problem, smoking cessation researchers are introducing relapse prevention techniques during pregnancy. They are also looking at the positive benefits associated with smoking as a way of developing sustainable cessation strategies. Since women who resume smoking cite its effectiveness in reducing stress and depression, cessation researchers are devising multi-intervention approaches to address those conditions. Interventions are woman-centered and carefully designed to avoid the rebound effect that sometimes ensues when fetal health is overemphasized.
Dr. Woodby observed that results from the self-help SCRIPP program underscore the importance of providing tailored interventions to pregnant women. Their experience also demonstrates the wisdom of gaining support from the health care provider community prior to program implementation and of establishing project advisory committees at each site.
Smoking cessation researchers are as interested in promoting preconceptional health messages as alcohol researchers are. Likewise, they are intrigued by the potential preventive role of the smoking woman’s partner although they are unclear about how best to involve social adjuncts in smoking cessation interventions. When specifically targeted to pregnant women, smoking cessation programs have found that self-help and brief interventions work well across cultural groups even though materials are not culturally specific. In Dr. Mullen’s opinion, the experience of pregnancy is the great leveler. However, it should be noted that differences in reading level and social class were taken into account in designing program materials. With respect to brief intervention, Dr. Mullen emphasized that it was more successful with lighter than heavier smokers. More work is needed to discern whether this approach is applicable to more dependent smokers.
Nondisclosure remains a problem. Although smoking cessation programs can take advantage of technology to check urine and saliva for cotinine, such tests do not currently exist for alcohol. Much like urinalysis for drugs, cutoffs overcome excuses related to passive smoking and supply providers with sound evidence of compliance. Smoking cessation researchers are also deliberately building harm reduction strategies into their interventions. For example, in the SCRIPP program, women who cannot quit make a contract to reduce their intake by half, an approach that appears to be effective.
For smoking cessation programs as for risk-drinking interventions, the most severely affected women pose the biggest challenge. For this reason, there is renewed interest in assessing the use of pharmacological adjuncts with pregnant women. Differences in providers’ abilities to reach pregnant women and motivate them to change are also being carefully evaluated. At the University of Houston’s programs, usual care providers are not typically involved. Instead, case managers adroit in responding to pregnant women’s multiple needs conduct the intervention, which includes stress reduction and help in resolving family conflicts. At SCRIPP sites, program personnel decide who will conduct the intervention. Since staff do not have time to explore fully smoking cessation as part of routine care, SCRIPP uses a trained person to introduce the self-help manual and an accompanying video. Followup is provided at a subsequent appointment that includes a 5-minute counseling session. Commenting on the CDC’s experience, Dr. Floyd said that the interventions delivered as part of their multisite trials were not intensive enough; one-on-one, face-to-face sessions with a dedicated provider were needed to make a difference. In Dr. Floyd’s opinion, this level and intensity of interaction would place an undue burden on existing providers.
Because so many women who smoke are also at-risk drinkers, Nancy Day asked the group how intervention programs can best coordinate multiple prevention messages. Although Dr. Mullen admitted that she was still uncertain about how best to approach this problem, she said that their case managers addressed risk drinking as one of many factors affecting pregnant women’s lives. Marc Usatin said that at Kaiser Permanente, the context in which the disclosure about drinking and/or smoking was made influenced the outcome. When a screening result is positive, the obstetrician or nurse conducting the screen literally escorts the woman to the staff member who will intervene. Their approach is both immediate and personal and seems to overcome a number of problems associated with disclosure and resistance.
- Many prevention strategies employed in smoking cessation programs probably can be applied to preventing risk drinking during pregnancy.
- Pregnancy-specific materials are more effective than generalized prevention materials with this population; referencing the shared experience of pregnancy overcomes cultural and other differences among target women.
- For women unable to quit smoking, harm reduction approaches have resulted in positive outcomes.
- Brief intervention works well across cultural groups but is more effective with lighter than heavier smokers.
- Nondisclosure of smoking (and at-risk drinking) behavior remains a serious problem.
- Better strategies are needed to reach women who smoke (and drink) heavily.