This study is consistent with previous research in which smokers who utilized cessation support in the past, when proactively reached and offered an opportunity to recycle in treatment, display high rates of treatment re-engagement [2
The major hypothesis of the current study was that proactive automated calls utilizing IVR technology would enhance recycling in treatment among Medicaid and uninsured former QL participants. This hypothesis received strong support as smokers reached by the IVR intervention were 11.2 times more likely to re-enroll in a new treatment cycle than the control group. This was an unexpectedly large odds ratio for the intervention effect, and the 95% confidence interval was also relatively large (5.4-23.3). The overall proportion of re-enrollment was 15% (78/521), while the unadjusted proportion of treatment arm smokers who re-enrolled was 28.2% (69/245), compared to the 3.3% (9/276) of control group smokers re-enrolling. As there was no evidence of limited variability among re-enrolling smokers on the logistic model independent variables (study arm, one or more chronic conditions, age, and propensity score), the most likely cause of the wide confidence interval is the relatively small cell size (N
9) for control group re-enrollees, which contributes uncertainty to OR estimation. Even with the relatively large confidence interval, the IVR intervention developed here – consisting of assessing barriers for recycling, delivering customized messages to address these barriers, and offering automatic transfer to the QL- was successful in recycling low income, adult smokers irrespective of age, gender, education, and race/ethnic background.
QLs often constitute the only source of professional support available for low income smokers, a segment of society that smokes at higher rates [15
] and reports lower success when attempting to quit than middle and high income smokers [14
]. Integration of IVR technology for re-enrollment procedures/interventions with QL cessation services can address a missed opportunity of reaching out to low income smokers who are satisfied with QL services and want to quit, but do not take the initiative to seek a new assisted quit attempt. Moreover, this can be done in a scalable way, considering that low income QL participants are difficult to reach (i.e., will not answer the telephone, have disconnected numbers on file, or hang up right after answering the call). IVR technology is a viable option to recycle smokers while minimizing the burden to QL staff of making numerous manual calls, thus allowing QL staff to dedicate their time to their primary task of providing smoking cessation counselling. Re-enrollment rates varied according to interest and confidence in quitting. While more than half (56.3%) of smokers reporting being ready to quit registered in QL services, 32% of those interested in quitting but reporting low confidence in their ability to succeed accepted a new cycle of treatment. It is possible that the delivery of a message normalizing relapse and encouraging a new quit attempt helped to achieve this relative high rate of treatment recycling in this group of smokers (see Additional file 1
for IVR message content). Three smokers (4.8%) who reported no interest in quitting chose to recycle back to QL support. This unexpected outcome could be related to a rapid change in interest in quitting, similar to what has been described as “unplanned quit attempts” reported in some population-based studies [21
]. Alternate explanations for these re-enrollments are that these callers pressed the wrong answer in the IVR system or were reluctant to report interest in quitting before learning more about what would be offered (in terms of pricing, services, or commitment).
Few smokers endorsed barriers to recycle related to QL services (eligibility, effectiveness or relevance). This finding was not surprising considering the study sample consisted of former QL participants, a service that is typically associated with high levels of satisfaction among its users [16
An important limitation of this study was our inability to reach 76% of participants from the original study sample, somewhat restricting the external validity of our findings. This reach rate is not uncommon in research with low income populations, who tend to switch and disconnect telephone services frequently. However, in this instance, the inability to reach participants was due mostly to calls never answered and not to disconnected numbers. We could have reached more people if our protocol implementation included the entire sample. Unfortunately, a programming error limited our reach by delivering 10 attempts to the primary phone to 699 participants, while the rest of the 2296 subjects in our sample received the full protocol (twenty attempts to the primary and secondary telephone numbers). Another factor that could boost participant reach in the future is the inclusion of caller identification, which would have assisted participants in making a decision on whether or not to take the call. In our study, we utilized a toll-free number not associated with a caller ID. In retrospect, we could have utilized a caller ID that identified the study, potentially increasing the chance of interested smokers accepting the call.
A second limitation relates to the timing of the intervention. Our sample was comprised of smokers who registered in QL support at least 12
months before the study was launched. This choice was made to meet the minimal period required by several states between first and new engagement into QL services. However, some studies suggest that proactive recycling recruitment should be attempted earlier. Fu and colleagues [3
] assessed interest among smokers who were treated for tobacco dependence in five Veterans Affairs medical centers. Almost two-thirds of relapsed smokers were interested in recycling into treatment within 30
days. Joseph and colleagues [5
] studied interest in further treatment among 2,340 smokers from the Minneapolis Veterans Administration Medical Center who received prescriptions for a smoking cessation aid during an 18-month period. Of continuing smokers, 98% were willing to make another quit attempt—50% immediately, and 28% within 1
month. In a population based recent study, Yeomans et al. [23
] found high variability on time elapsed since a new quit attempt among 427 current and former smokers, varying from 1–1162
days. By 6
months, the cumulative proportion of subjects making their next quit attempt was 24.5%, which increased to 52.9% by 12
months (no information is given on what the proportion of these new quit attempts were supported by treatment.). Future studies may want to compare different timing of proactive recycling recruitment of relapsed smokers.