Abstract
Objectives. We examined the coevolution of information sharing and implementation of evidence-based practices among US and Canadian tobacco cessation quitlines within the North American Quitline Consortium (NAQC).
Methods. Web-based surveys were used to collect data from key respondents representing each of 74 participating funders of NAQC quitlines during the summer and fall of 2009, 2010, and 2011. We used stochastic actor-based models to estimate changes in information sharing and practice implementation in the NAQC network.
Results. Funders were more likely to share information within their own country and with funders that contracted with the same service provider. Funders contracting with larger service providers shared less information but implemented significantly more practices. Funders connected to larger numbers of tobacco control researchers more often received information from other funders. Intensity of ties to the NAQC network administrative organization did not influence funders’ decisions to share information or implement practices.
Conclusions. Our findings show the importance of monitoring the NAQC network over time. We recommend increased cross-border information sharing and sharing of information between funders contracting with different and smaller service providers.
Tobacco use is still the leading cause of preventable death in Canada and the United States. Although smoking among adults has been declining, currently 16% of Canadian adults and 19% of American adults are smokers.1,2 Many smokers trying to quit fail, and those who succeed often have attempted to quit multiple times.3,4 A number of governments, including that of Canada, have required tobacco companies to include warning labels with toll-free numbers for tobacco cessation quitlines, and many, including the US government, have launched mass media campaigns promoting these toll-free numbers.
In North America, a quitline is a partnership between 2 or more organizations (composed of 1 or 2 funders and a provider organization) in which telephone-based tobacco cessation services are provided to people trying to quit smoking. The funder is typically the department of public health in each US state or a similar government agency in Canada. Funders determine the services to be offered while contracting out responsibility for actual service provision to public, nonprofit, or for-profit providers.
In the United States and Canada, 65 quitlines now operate independently and provide various cessation services. Even though substantial evidence is available regarding the specific cessation practices that are effective, not all quitlines implement the same practices.5 However, all quitlines have the same goal (helping smokers quit) and share information and resources regarding tobacco cessation services under the umbrella of the North American Quitline Consortium (NAQC). A network administrative organization (NAO)6 facilitates the coordination and governance of the NAQC network and offers professional support, educational programs, and activities designed to help improve quitlines.7
We examined the evolution of information sharing and implementation of evidence-based practices in this network; our specific focus was on funders given that they ultimately determine the services to be offered by their quitline. Figure 1 shows all of the NAQC organizations and the service delivery information ties reported by funders in 2009.
FIGURE 1—
Service delivery information received and reported by funder organizations in 2009: North American Quitline Consortium.
Note. F = funder; NAO = network administrative organization; P = service provider; QL = quitline.
aNonrespondent.
Networks have been found to play an important role in the dissemination and provision of evidence-based practices in health and human services.8–12 Organizations often depend on information from others for determining the practices that are most effective,5,13–15 and teasing out which sources influence service delivery is important for public health networks aiming to disseminate information to improve service delivery. Knowing the factors that condition the selection of information sources and which of these sources influence the implementation of evidence-based practices will increase our understanding of the dynamics underlying dissemination of practices within networks.
Previous studies showed that awareness of evidence-based practices was highest among NAQC funders that were strongly connected to researchers and that greater network centralization through the consortium’s NAO was associated with higher adoption rates.5,11 However, these studies did not focus on implementation of evidence-based practices and were unable to draw causal conclusions. Information sharing among funders and implementation of services may be influenced by the local network in which a funder is embedded. Funders may depend on other funders, providers, the NAO, and researchers to receive information regarding the efficacy, reach, and delivery of tobacco cessation services, and by doing so they potentially increase the effectiveness of their quitlines. Along the lines of a homophily argument, funder-to-funder ties may be most informative and influential. However, diversity of information sources might lessen the dependence on funder ties.16 If a funder is strongly connected to the NAO, service providers, or researchers, implementation of evidence-based practices may increase but connections to other funders may become redundant.
Factors such as funder country, quitline age, and provider size may also play a role in network and implementation dynamics. Network ties are often bounded by geospatial conditions,17,18 and US and Canadian quitlines show a tendency to share information within their own country.7 Older quitlines may be more active or successful in establishing and maintaining network links,19 consistent with an accumulative advantage argument.16 Furthermore, because a provider may have contractual ties with funders of multiple quitlines, it can be expected that a funder contracting with such a provider will learn from the other funders working with that provider, which may influence implementation of practices.
Our goals were to develop a network analysis model of public health service delivery and to provide insight into the functioning of tobacco quitlines as a specific case study. We used stochastic actor-based models20 to simultaneously examine the evolution of information sharing among funders and implementation of evidence-based practices by funders’ quitlines. These advanced network models can represent an extensive assortment of influences on network change, allow estimation of parameters expressing these influences, and provide tests of corresponding hypotheses.21 Such models have been used successfully in a number of studies to examine the evolution of various organizational networks.19,22–25 However, none of these investigations focused on information sharing and simultaneously modeled the evolution of an outcome.
Our study can serve as an example of the ways in which public health services in the United States and Canada are organized, with governments contracting for health services with providers or engaging with the research community to address public health issues. Our 2 primary research questions were as follows: Which network ties and funder characteristics influence information sharing among funders? and Which network ties and funder characteristics influence implementation of evidence-based practices?
METHODS
We used a Web-based survey to collect longitudinal data from all NAQC organizations during the summer and early fall of 2009 (time 1 [T1]), 2010 (T2), and 2011 (T3). Depending on organization size, data were collected from 1 to 5 respondents. During recruitment, each primary contact listed within the NAQC database was asked to submit a list of all individuals within his or her organization who played a key role in decision making related to implementation of new practices. This list was updated at each observation point, but every attempt was made to contact the same individuals at each point. To reduce attrition, we made follow-up contacts with respondents via e-mail and telephone.7
Our study focused on the network of funders. Overall, 65 of 72 funders (representing 60 of 63 quitlines) responded in 2009, and 66 of 74 (representing 64 of 66 quitlines) responded in 2010 and 2011 (additional information on respondent attrition is provided in Appendix A, available as a supplement to the online version of this article at http://www.ajph.org). Questions and methods were pretested with a working group of network members who agreed to provide initial feedback.
Funders’ Network Ties
Each year of the study, funder respondents received a list of all existing quitline funders as well as all service providers (n = 22) and were asked to indicate whether they had received service delivery information from them (e.g., ideas about improving or assessing treatments or services, new protocols, or new treatment technologies) and at what level of intensity (0 = no involvement, 1 = low, 2 = moderate, 3 = high). Only moderate- and high-intensity ties were included in our analyses.
To measure the intensity of ties to the consortium’s NAO, we asked all respondents to indicate at what level of intensity they received service delivery information from the NAO at T1 and T2 (0 = no involvement, 1 = low, 2 = moderate, 3 = high). Finally, as a means of assessing numbers of ties to researchers, respondents were asked at T1 and T2 to indicate whether they had received information from each tobacco control researcher included on a list that had been provided to them.
Because a single individual could establish a tie without the involvement of others,26 a relationship was classified as being present if any of the respondents in the organization indicated that a tie existed. If a quitline had 2 funders, data were aggregated in the same way. As a result, quitlines were the unit of analysis in our funder network data.
Implementation of Evidence-Based Practices
To measure implementation of evidence-based practices within quitlines, a working group of practitioners and researchers involved with the NAQC developed a list of 23 practices that were believed to be most effective in attracting smokers to receive help and encouraging them to reduce their smoking or quit. Each year, for each quitline’s funder and service provider organization, one key responder most likely to have the most complete up-to-date information regarding the implementation of these practices was identified and asked to indicate whether the quitline had fully implemented the practices. Because there were certain practices about which funders were more likely to have accurate, up-to-date knowledge and other practices about which providers were more likely to be knowledgeable, we used only the information from the funder key responders for practices in the funder domain and only the information from the provider key responders for practices in the provider domain (see Appendix B, available as a supplement to the online version of this article at http://www.ajph.org).
The evidence-based practices included in our analyses are shown in the box on the previous page. We excluded 2 practices because they were applicable only to the United States, which could have biased our results. We also excluded a pair of other practices, one because it overlapped completely with a combination of 2 other already-included practices and another owing to an absence of any significant scientific evidence that it indeed would be effective (see the box on the previous page.). In the case of all quitlines, we determined how many of the remaining 19 practices were fully implemented at each observation.
Evidence-Based Smoking Cessation Practices Included in the Study Analysis
| 1. Evaluate the effectiveness of the quitline |
| 2. Conduct mass media promotions for targeted populations |
| 3. Conduct mass media promotions for the mainstream population |
| 4. Fax-to-quit or fax-referral programa |
| 5. Integrate telephone counseling with face-to-face cessation services through referrals or combinations of services |
| 6. Integrate telephone counseling with Web-based, Internet-based, or eHealth programs through referrals or combinations of services |
| 7. Provide a multiple call protocol (2 or more calls related to the same quit attempt) |
| 8. Provide counseling immediately to all callers who request it |
| 9. Provide free (or discounted) nicotine replacement therapy to callers without requiring counseling |
| 10. Provide nicotine replacement therapy to callers but require that they register for counseling |
| 11. Provide proactive (outbound) telephone counseling |
| 12. Provide reactive (inbound) counseling |
| 13. Provide self-help materials for tobacco users who receive counseling |
| 14. Provide self-help materials to proxy callersb |
| 15. Recontact relapsed smokers for reenrollment in quitline services |
| 16. Refer callers with insurance to health plans that provide telephone counseling |
| 17. Supplement quitline services with interactive voice response services |
| 18. Train provider groups on “2As” (ask, advise) or “3As” (ask, advise, assist) and referral of tobacco users |
| 19. Use text messaging to provide tailored support in conjunction with or instead of telephone counseling |
Note. Two practices (obtaining Medicaid reimbursement or other types of reimbursement for counseling provided to callers and serving callers without insurance coverage) were not included in the analysis because they occurred only in the United States. One practice (providing self-help materials for tobacco users regardless of their reason for calling or the services selected) was excluded because of its complete overlap with practices 13 and 14 (in combination). Finally, 1 practice (staffing the quitline with counselors who meet or exceed the requirement of master’s-level training) was excluded owing to the lack of scientific evidence that it is effective.
A quitline receives a tobacco user's contact information from a health care provider by fax and makes proactive outbound call attempts to reach him or her and offer enrollment in the quitline.
Proxy callers are individuals other than tobacco users (a family member, spouse, or friend) interested in information on quitting tobacco so they can help the user quit more effectively.
Other Data
The consortium’s NAO provided data on provider size at T1 and T2. These data reflected the number of quitlines with which a provider shared a contractual tie to offer smoking cessation services (grouped into 5 categories: 1 = 1, 2 = 2–5, 3 = 6–10, 4 = 11–15, 5 = more than 15), the number of years the quitline had been in operation, and the country of the quitline (0 = United States, 1 = Canada).
Data Analysis
The unconditional method of moments20,27 in RSiena28 was used to analyze a stochastic actor-based coevolution model. Details on (mathematical) specification of actor-based models are provided elsewhere.20,21
The actor-based coevolution model we used is a combination of 2 models: the network evolution model and the implementation evolution model. The network evolution model specifies the preferred direction of network change via a list of effects measuring network structure (endogenous effects) and funder attributes (exogenous effects). Two effects are included for each funder attribute: ego effects and alter effects. A positive ego effect indicates that funders scoring higher on a given attribute have a greater tendency to receive information from other funders, whereas a positive alter effect indicates that funders are more likely to receive information from funders scoring high on that attribute than funders with low scores. To control for funders’ tendency to receive information from other funders contracting with the same provider or funders within the same country, we included “same-provider” and “same-country” effects (the complete network evolution portion of the combined model is described in Appendix C, available as a supplement to the online version of this article at http://www.ajph.org).
The implementation evolution model specifies directions of change in the implementation of evidence-based practices according to a list of network, behavior, and other attribute effects on which probabilities of changes in implementation may depend. (The complete list is provided in Appendix C.)
The effects we included were tested via t ratios (estimates divided by standard errors) with approximately standard normal null distributions.27 Funders entering the study after the initial time point were included for the duration of their membership in the network.29 Alter squared effects (controlling for the possibility that some attribute effects were nonlinear in their evolution), interaction effects between ego and alter terms (modeling whether receipt of information is based on attribute similarities), and higher-order degree-related effects (explained elsewhere28) were assessed by means of score tests30 and included only if they were significant. Also, we added time-changing variables to our model in cases of time heterogeneity.28 Finally, we calculated goodness of fit statistics.31 A model’s goodness of fit is considered acceptable if P values are above .05 (see Appendix D, available as a supplement to the online version of this article at http://www.ajph.org).
RESULTS
Table 1 presents the number of different network organizations examined and shows overall and individual funder (network) characteristics over time. The average number of reported outgoing ties (out degree) ties increased during the first study period but decreased during the second period. The average number of out-degree ties over time was 2.84. The average number of fully implemented practices increased over time, from 9.3 at T1 to 11.24 at T3. Figure 2 shows implementation rates for each practice.
TABLE 1—
North American Quitline Consortium (NAQC) and Funder Network Characteristics, 2009–2011
| Variable | Time 1 (2009) | Time 2 (2010) | Time 3 (2011) |
| NAQC organizations/individuals, no. | |||
| Quitlines | 60 | 64 | 64 |
| Funders | 65 | 66 | 66 |
| Funder respondents | 122 | 136 | 105 |
| Service providers | 22 | 22 | |
| NAO | 1 | 1 | |
| NAQC funder network structure | |||
| No. of ties, mean | 3.00 | 3.48 | 2.05 |
| Densitya | 0.05 | 0.06 | 0.03 |
| Isolated funders,b no. | 6 | 8 | 16 |
| Reciprocity indexc | 0.15 | 0.15 | 0.12 |
| Jaccard indexd | 0.22 | 0.21 | |
| Individual attributes of funders, mean (SD; range) | |||
| No. of practices implemented | 9.30 (2.86; 3–16) | 10.00 (2.41; 4–15) | 11.24 (2.59; 6–17) |
| Intensity of ties to NAO | 2.08 (0.72; 1–3) | 2.14 (1.02; 0–3) | |
| No. of provider ties | 0.83 (1.42; 0–6) | 0.76 (1.19; 0–6) | |
| No. of researcher ties | 3.61 (2.71; 0–11) | 4.29 (3.19; 0–13) | |
| Provider sizee | 3.19 (1.50; 1–5) | 3.29 (1.66; 1–5) | |
| Quitline age, y | 6.43 (2.97; 1–16.5) |
Note. NAO = network administrative organization. A total of 15.6% of funders were Canadian.
Number of information-receiving ties as a fraction of all possible ties.
Funders that have no incoming or outgoing ties.
Proportion of ties that are mutual.
A measure of the amount of change between 2 network data collection points. Values below 0.2 raise doubts about the applicability of an actor-based model.32
Provider size was coded as follows: 1 = only 1 contractual tie, 2 = 2–5 ties, 3 = 6–10 ties, 4 = 11–15 ties, 5 = more than 15 ties.
FIGURE 2—
Numbers of quitlines that fully implemented smoking cessation practices: North American Quitline Consortium, 2009–2011.
Note. 2As = ask, advise; 3As = ask, advise, assist; NRT = nicotine replacement therapy. For details on practices, see the box on page 1816.
Network evolution model results are reported in Table 2. Funders tended not to receive information from other arbitrary funders (without consideration of specific attributes of those funders). Funders did receive information from funders that already received information from them (reciprocity). A significant transitive ties effect implied that funders received information from funders they already received information from (if A received information from B and B in turn received information from C, A would have a significant tendency to receive information from C).
TABLE 2—
Results of the Actor-Based Model: North American Quitline Consortium Network, 2009–2011
| b (SE) | P | |
| Network evolution | ||
| Rate: period 1a | 19.283 (3.901) | |
| Rate: period 2a | 12.838 (1.794) | |
| Out-degree (outgoing ties) | −5.091*** (0.449) | |
| Reciprocity | 0.874*** (0.243) | |
| 3 cyclesb | −0.008 (0.129) | |
| Transitive ties | 0.587* (0.200) | |
| In-degree popularity (square root) | 0.335* (0.092) | |
| In-degree activity (square root) | −0.217 (0.202) | |
| Out-degree activity (square root) | 0.384*** (0.065) | |
| Implementation egoc | 0.055 (0.035) | |
| Implementation alterc | 0.047 (0.035) | |
| Intensity of tie to NAO ego | 0.014 (0.068) | |
| Intensity of tie to NAO alter | 0.046 (0.065) | |
| No. of provider ties ego | −0.055 (0.039) | |
| No. of provider ties alter | −0.013 (0.032) | |
| No. of researcher ties ego | 0.074** (0.026) | |
| No. of researcher ties alter | 0.016 (0.022) | |
| Age ego | −0.007 (0.026) | |
| Age alter | 0.060** (0.020) | |
| Country ego | 0.135 (0.248) | |
| Country alter | 0.442 (0.293) | |
| Same country | 1.413*** (0.287) | |
| Provider size ego | −0.087* (0.038) | |
| Provider size alter | −0.162** (0.050) | |
| Provider size squared alter | −0.070* (0.030) | |
| Same provider | 0.610*** (0.124) | |
| Time dummy: out-degree activity | −0.168*** (0.041) | |
| Time dummy: provider size ego | −0.080 (0.071) | |
| Implementation behavior evolution | ||
| Rate: period 1a | 8.194 (2.754) | |
| Rate: period 2a | 19.241 (11.823) | |
| Implementation linear shape | 0.257 (0.467) | |
| Implementation quadratic shape | −0.065 (0.044) | |
| In-degree (incoming ties) | −0.021 (0.039) | |
| Out-degree (outgoing ties) | −0.028 (0.030) | |
| Isolation | −0.419 (0.748) | |
| Popularity alter | 0.021 (0.052) | |
| Implementation similarity | 0.082 (0.855) | |
| Intensity of tie to NAO | 0.077 (0.060) | |
| No. of provider ties | 0.008 (0.040) | |
| No. of researcher ties | −0.005 (0.021) | |
| Age | 0.025 (0.022) | |
| Country | 0.044 (0.169) | |
| Provider size | 0.114* (0.051) | |
| Time dummy: out-degree | 0.027 (0.034) | |
| Goodness of fitd | ||
| Out-degree distribution | .055 | |
| Behavior distribution | .944 | |
| In-degree distribution | .343 | |
| Geodesic distribution | .171 | |
| Triad census | .43 |
Note. NAO = network administrative organization.
Rate parameters, calculated separately for each period, account for the speed at which the network changes and practice implementation changes between 2 subsequent observations.
This effect refers to generalized reciprocity or the opposite of hierarchy in the network. Funder A receives information from funder B, funder B from funder C, and funder C from funder A.
Ego effects model the impact of a funder’s own attributes on whether the funder (ego) will receive information from other funders (alters). Alter effects model the impact of attributes of other funders (alters) on whether a funder (ego) will receive information from them.
More information on goodness of fit statistics is provided in Appendix D (available as a supplement to the online version of this article at http://www.ajph.org).
*P < .05; **P < .01; ***P < .001.
In addition, there was a significant tendency for funders to receive information from funders many other funders received information from (in-degree popularity effect). Relative to funders receiving information from few other funders (low out degree), funders receiving information from many other funders (high out degree) exhibited a higher tendency to receive information from even more funders (out-degree activity effect). However, this tendency decreased significantly over time, as shown by the significant negative out-degree activity time dummy effect.
Funders connected to increased numbers of researchers received information from more funders. Funders that contracted with larger providers had a significantly lower tendency to receive information from other funders and were less frequently a source of information. Funders were more likely to receive information from funders of older quitlines. In addition, funders of older quitlines received information from fewer funders than funders of younger quitlines. Finally, there was a significant tendency for funders to receive information from funders that contracted with the same provider and from funders within their own country.
Implementation evolution model results are also reported in Table 2. Only provider size influenced practice implementation. Funders contracting with larger providers tended to implement more practices than funders contracting with smaller providers. The goodness of fit of the model was acceptable for all 5 network indices (out-degree distribution, behavior distribution, in-degree distribution, geodesic distribution, triad census; see Appendix D).
DISCUSSION
Our aim in this study was to examine the coevolution of information-receiving ties among NAQC funders and quitlines’ implementation of evidence-based practices. The average number of implemented practices increased over time. The number of reported ties, however, decreased during the second study period. Although this decrease may have been due to survey fatigue, it is possible that tobacco control organizations struggled with funding, possibly leading to decreases in staffing, reductions in the amount of time available to interact with other funders, or a focus on the most useful ties and a discontinuation of unimportant ones.
We assessed which funder characteristics and information-receiving ties influenced receipt of information among funders. In line with previous studies focusing on communication and advice networks among organizations,23–25 funders had a significant tendency to return information to each other. Also similar to earlier studies on advice networks,23,24 funders tended to receive information from funders that were tied to one of their other advisors, as well as funders from which many other funders received information.
Our results also showed that funders contracting with larger providers shared information with fewer funders. It is likely that funders contracting with larger providers can gain knowledge indirectly from those providers’ other contractors and therefore depend less on obtaining information from other funders. The important role of providers is also reflected by the significant tendency of funders to receive information from other funders contracting with the same providers. In all likelihood, funders learn about evidence-based practices from their providers. Given that providers may consider some information about quitline practices proprietary, it might be easier for funders to obtain information from other funders that contract with their provider. In addition, information obtained from other funders with the same provider may be perceived as more relevant because the systems and infrastructures used to provide services are the same. A provider will often make a new practice available to all of its quitlines’ funders at the same time, increasing opportunities for these funders to share information.
Ties to researchers also played an important role in the evolution of information-receiving ties among funders. Funders well connected to researchers more often received information from other funders. Indirectly, researchers may stimulate information sharing among funders. Because researchers are often aware of funders doing innovative work, other funders often ask them for introductions or referrals to these organizations. Researchers may also be interested in a specific area of investigation. Funders discussing possible partnerships may share information about related practices.
In line with previous research examining the effects of organizational age on establishing and maintaining ties within a network,19 funders from younger quitlines received information from more funders, whereas information was more likely to be received from funders of older quitlines. This suggests that experience is valuable and shared from older to younger quitlines.
Finally, in line with previous studies focusing on the NAQC,7,11 there was a tendency for funders to share information with other funders within their own country. Funders may believe that their experiences are country specific and that what is effective in the United States is not necessarily effective in Canada (and vice versa). US quitlines tend to be better funded than Canadian quitlines, and US funders tend to play a more active role in decision making regarding service delivery. Moreover, each country has its own tobacco control conference, limiting cross-country interactions.
In the context of our study, and considering the variables we assessed, our results showed that practice implementation was influenced only by provider size. Funders contracting with a larger provider had a significantly higher tendency than funders contracting with a smaller provider to implement evidence-based practices. This result might be explained by the fact that providers often offer funders an initial list of practices from which to choose. If funders want to implement a practice that is not offered by their current provider, they may change providers. Once providers begin offering the practice, many of their funders may start to implement it. Furthermore, larger providers can gain knowledge from a larger number of quitlines and have more resources to work with, allowing them to expand the range of services offered.
Ties to researchers did not influence the actual implementation of evidence-based practices according to our findings. However, in earlier cross-sectional research, it was concluded that funders well connected to researchers were more likely to be aware of evidence-based practices.5 It is feasible that researchers can influence funders’ awareness of evidence-based practices but not implementation of these practices. Although tobacco control researchers are in the best position to recommend practices that are effective in helping smokers quit, they may have less practical knowledge regarding the implementation process.
Although our findings indicated that the intensity of ties with the NAQC’s NAO does not seem to influence the number of evidence-based practices implemented or information sharing among funders, the NAO is a central component of the service delivery information-sharing network11 and may still play an important role in dissemination of knowledge. The NAO seeks to provide information to quitlines that they can use to make decisions about whether to implement (or remove) practices on the basis of emerging evidence and many other factors that must be considered. Furthermore, the NAO can reach out to quitline organizations that may not be well positioned in the flow of information among funders.
Limitations
Our study involved a number of limitations. First, we used self-reported data. Although confidentiality was assured, some respondents may have overreported connections to other organizations to represent their organization favorably. Second, differences between moderate-intensity and high-intensity ties were not a focus in this study. Future research should examine whether our findings would differ among funders with moderate-intensity and those with high-intensity ties. Third, we assessed only service delivery information-receiving ties. Although these ties are very important, funders also share information on financial or managerial matters, such as budgets or staffing, that could have an impact on implementation processes.
Finally, to analyze the coevolution of the funder network and implementation of practices, we calculated the total number of fully implemented evidence-based practices. It is feasible that different processes are involved in the early stages of implementation and in different evidence-based practices. Furthermore, to avoid biased results, we excluded 2 practices—serving individuals without insurance and obtaining Medicaid reimbursement or other types of reimbursement for counseling provided to callers—because they applied only to US quitlines. In the United States, these practices are very important for low-income smokers without insurance. Separate analyses of each specific practice implemented throughout the complete network or the network of US quitlines only seem warranted.
Conclusions
Our findings imply that it is important to continue to monitor the NAQC network and encourage quitline organizations to share knowledge. The consortium’s NAO seems to have assisted in fostering an environment of information sharing. However, stimulation of cross-border communication could help funders in the 2 countries learn from each other. Because providers seem to play an important role in the information network, and because funders contracting with larger providers implemented significantly more evidence-based practices, funders in general could benefit from sharing information with funders outside of their provider cluster, especially those contracting with smaller providers.
Our results can be of relevance beyond the quitline community. Our methods and outcomes are relevant to public health issues involving networks and evolving health care systems; specifically, they are applicable to cases in which governments contract for health services with providers or engage with researchers in addressing health issues. Our findings point to the need for research designed to provide a better understanding of how collaborative care networks can be optimized to provide the greatest service quality.
Acknowledgments
This study was conducted as part of the KIQNIC project (Knowledge Integration in Quitlines: Networks That Improve Cessation), funded by a grant (R01CA128638-04) from the National Institutes of Health to the University of Arizona and the Mayo Clinic. Additional support was provided by a National Cancer Institute Cancer Center Support Grant (CCSG-CA 023074).
We thank Joseph Bonito, Gregg Moor, and Jonathan Beagles for their valuable input and assistance and all of the quitline organizations and their representatives for their willingness to participate and commitment. Special thanks to our friend and colleague Keith Provan, who unexpectedly passed away during the writing of this article, for all of his help and guidance.
Human Participant Protection
This study was approved by the University of Arizona Human Subjects Protection Program. Participants were e-mailed full study disclosure information, and it was explained to them that by taking part in the survey they were giving permission for their information to be used for research purposes.
References
- 1.Centers for Disease Control and Prevention. Vital signs: current cigarette smoking among adults aged ≥18 years—United States, 2005–2010. Morbidity and Mortality Weekly Report. 2011;60(35):1207–1212. [PubMed] [Google Scholar]
- 2.Health Canada. Canadian Tobacco Use Monitoring Survey, 2012. Available at: http://www.hc-sc.gc.ca/hc-ps/tobac-tabac/research-recherche/stat/_ctums-esutc_2012/ann_summary-sommaire-eng.php. Accessed March 12, 2015.
- 3.Cohen S, Lichtenstein E, Prochaska JO et al. Debunking myths about self-quitting: evidence from 10 prospective studies of persons who attempt to quit smoking by themselves. Am Psychol. 1989;44(11):1355–1365. doi: 10.1037//0003-066x.44.11.1355. [DOI] [PubMed] [Google Scholar]
- 4.Messer K, Trinidad DR, Al-Delaimy WK, Pierce JP. Smoking cessation rates in the United States: a comparison of young adult and older smokers. Am J Public Health. 2008;98(2):317–322. doi: 10.2105/AJPH.2007.112060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Provan KG, Beagles JE, Mercken L, Leischow SJ. Awareness of evidence-based practices by organizations in a publicly funded smoking cessation network. J Public Adm Res Theory. 2013;23(1):133–153. doi: 10.1093/jopart/mus011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Provan KG, Kenis P. Modes of network governance: structure, management, and effectiveness. J Public Adm Res Theory. 2008;18(2):229–252. [Google Scholar]
- 7.Leischow SJ, Keith P, Jonathan B et al. Mapping tobacco quitlines in North America: signaling pathways to improve treatment. Am J Public Health. 2012;102(11):2123–2128. doi: 10.2105/AJPH.2011.300529. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fitzgerald L, Ferlie E, Wood M, Hawkins C. Interlocking interactions, the diffusion of innovations in health care. Hum Relations. 2002;55(12):1429–1449. [Google Scholar]
- 9.Fujimoto K, Valente TW, Pentz MA. Network structural influences on the adoption of evidence-based prevention in communities. J Community Psychol. 2009;37(7):830–845. doi: 10.1002/jcop.20333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Provan KG, Lemaire RH. Core concepts and key ideas for understanding public sector organizational networks: using research to inform scholarship and practice. Public Adm Rev. 2012;72(5):638–648. [Google Scholar]
- 11.Lemaire RH, Provan KG, Leischow SJ. Getting people to quit smoking: the evolution of information sharing ties in the North American Quitline Consortium. Available at: http://www.union.wisc.edu/pmra2013/Paper%20Submissions/Renamed/Getting%20People%20to%20Quit%20Smoking%20The%20Evolution%20of%20Information%20Sharing%20Ties%20in%20the%20North%20American%20Quitline%20Consortium.pdf. Accessed March 12, 2015.
- 12.Valente TW, Chou CP, Pentz MA. Community coalitions as a system: effects of network change on adoption of evidence-based substance abuse prevention. Am J Public Health. 2007;97(5):880–886. doi: 10.2105/AJPH.2005.063644. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Boaz A, Pawson R. The perilous road from evidence to policy: five journeys compared. J Soc Policy. 2005;34(2):175–194. [Google Scholar]
- 14.Jennings ET, Hall JL. Evidence-based practice and the use of information in state agency decision making. J Public Adm Res Theory. 2012;22(2):245–266. [Google Scholar]
- 15.Harris JK, Luke DA, Burke RC, Mueller NB. Seeing the forest and the trees: using network analysis to develop an organizational blueprint of state tobacco control systems. Soc Sci Med. 2008;67(11):1669–1678. doi: 10.1016/j.socscimed.2008.07.013. [DOI] [PubMed] [Google Scholar]
- 16.Powell WW, White DR, Koput KW, Owen-Smit J. Network dynamics and field evolution: the growth of inter-organizational collaboration in the life sciences. Am J Sociol. 2005;110(4):1132–1205. [Google Scholar]
- 17.Bell GG, Zaheer A. Geography, networks, and knowledge flow. Organ Sci. 2007;18(6):955–972. [Google Scholar]
- 18.Graddy EA, Chen B. Influences on the size and scope of networks for social service delivery. J Public Adm Res Theory. 2006;16(4):533–552. [Google Scholar]
- 19.Margolin DB, Shen C, Lee S, Weber MS, Fulk J, Monge P. Normative influences on network structure in the evolution of the children’s rights NGO network, 1977–2004. Communication Res. 2015;42(1):30–59. [Google Scholar]
- 20.Snijders TAB, Steglich C, Schweinberger M. Modeling the co-evolution of networks and behavior. In: Van Montfort K, Oud H, Satorra A, editors. Longitudinal Models in the Behavioral and Related Sciences. Mahwah, NJ: Lawrence Erlbaum Associates; 2007. pp. 41–71. [Google Scholar]
- 21.Snijders TAB, van de Bunt GG, Steglich CEG. Introduction to stochastic actor-based models for network dynamics. Soc Networks. 2010;32(1):44–60. [Google Scholar]
- 22.van de Bunt GG, Groenewegen P. An actor-oriented dynamic network approach: the case of interorganizational network evolution. Organ Res Meth. 2007;10(3):463–482. [Google Scholar]
- 23.Agneessens F, Wittek R. Where do intra-organizational advice relations come from? The role of informal status and social capital in social exchange. Soc Networks. 2012;34(3):333–345. [Google Scholar]
- 24.Lazega E, Mounier L, Snijders T, Tubaro P. Norms, status and the dynamics of advice networks: a case study. Soc Networks. 2012;34(3):323–332. [Google Scholar]
- 25.Walker KL, Stohl C. Communicating in a collaborating group: a longitudinal network analysis. Commun Monogr. 2012;79(4):448–474. [Google Scholar]
- 26.Maurer I, Ebers M. Dynamics of social capital and their performance implications: lessons from biotechnology start-ups. Adm Sci Q. 2006;51(2):262–292. [Google Scholar]
- 27.Snijders TAB. The statistical evaluation of social network dynamics. Sociol Methodol. 2001;31(1):361–395. [Google Scholar]
- 28.Ripley RM, Snijders TAB, Boda Z, Voros A, Preciado P. Manual for RSiena. Available at: http://www.stats.ox.ac.uk/∼snijders/siena/RSiena_Manual.pdf. Accessed March 12, 2015.
- 29.Huisman M, Snijders TAB. Statistical analysis of longitudinal network data with changing composition. Sociol Methods Res. 2003;32(2):253–287. [Google Scholar]
- 30.Schweinberger M. Statistical modeling of network panel data: goodness-of-fit. Available at: http://sites.stat.psu.edu/∼mus47/publications/goodness-of-fit.pdf. Accessed March 12, 2015. [DOI] [PubMed]
- 31. Lospinoso JA, Snijders TAB. Goodness of fit for stochastic actor oriented models. Paper presented at: Sunbelt XXXI; February 2011; St. Pete Beach, FL.
- 32.Veenstra R, Steglich CEG. Actor-based model for network and behavior dynamics. In: Laursen B, Little TD, Card NA, editors. Handbook of Developmental Research Methods. New York, NY: Guilford Press; 2011. pp. 598–618. [Google Scholar]