JOURNAL OF TROPICAL PEDIATRICS, VOL. 59, NO. 5, 2013 Brief Report Effect of Households Social Networks on Lice Infestation among Vulnerable Mexican Children: a Qualitative Comparative Analysis by Lydia Ortega-Marı n, 1 Margarita Márquez-Serrano, 2 Luz M. Lara-Lo pez, 3 Ligia I. Moncada, 4 and Alvaro J. Idrovo 5 1 General Direction of Health Promotion, Sub-Ministry of Prevention & Promotion, Ministry of Health. Mexico City, Mexico 2 Centre for Health System Research, National Institute of Public Health. Cuernavaca, Morelos, Mexico 3 Academic Secretary, National Institute of Public Health. Cuernavaca, Morelos, Mexico 4 Public & Tropical Health Department, School of Medicine, National University of Colombia. Bogota, DC, Colombia 5 Cardiovascular Foundation of Colombia. Floridablanca, Santander, Colombia, and Public Health Department, School of Medicine, Industrial University of Santander. Bucaramanga, Santander, Colombia Correspondence: Margarita Ma rquez-serrano. E-mail:<margarita.marquez@insp.mx>. Correspondence address: Avenida Universidad 655. Colonia Santa Marıá Ahuacatitla n, CP. 62100. Cuernavaca, Morelos, Mexico. Summary The prevalence of pediculosis is high among elementary and secondary school children, which favors the belief that infestation occurs more often in schools than in homes. This study explored the role of households social networks in the transmission of head lice. Seventeen school children and their social networks (n ¼ 22) from Acatlipa (Morelos, Mexico) participated in a prospective observational study during school vacation. The hair of all the school children was washed with shampoo containing permethrin at the beginning of the study and the incidence of pediculosis (O) was evaluated at the beginning of the school term (follow-up at 1.5 months). The sets included in the qualitative comparative analysis were sex (S), length of hair (H), baseline diagnostic of pediculosis (I) and degree (D) and infestation index (N) obtained through the analysis of social networks. The prevalence of pediculosis was the same at the beginning and the end of follow-up (17.6%). The degree of the school children s networks ranged between 2 and 14. There were 8 configurations, the most frequent being F*i*d*n*h. The most parsimonious configuration associated with the incidence of pediculosis was F*I*d*H (female, previous infestation, low degree and long hair), with a coverage of 0.344 and a consistency of 0.941. Indicators of social networks made it possible to identify the role of households social networks in the transmission of lice. Individual actions such as the use of shampoo containing insecticides are temporary and, therefore, structural actions should be favored. Key words: lice infestation, school children, social network analysis, neglected diseases, poverty, Mexico. Introduction Infestation of head lice is produced by Pediculus humanus var. capitis. Because this is not usually Acknowledgements We would like to thank the participants in the study and teachers for their collaboration during the field work. This study was supported by the Mexican National Institute of Public Health. included as a priority for health programs, there are few efforts to control it, such that it is considered to be a neglected parasitism [1]. Prevalence of infestation is heterogeneous among nations, populations and cultures (0.7 61.4%). However, higher prevalences are frequent among school children, girls and women [2]. In Mexico, there are only three studies on pediculosis, and prevalences were ranging between 18 and 60% [3]. The belief that school is the primary site of contagion has become widespread, minimizing the role that the household may play in transmission and favoring the stigmatization of infested individuals. ß The Author [2013]. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com 413 doi:10.1093/tropej/fmt041 Advance Access published on 10 June 2013
Infested children are frequently not permitted to attend school even though it is known that this measure is not adequate for controlling pediculosis [4]. Household is a propitious site of contagion because it is where the children interact closely with their families, facilitating transmission by head head contact, which is known to be the primary means of infestation [5]. Lice infestation is recognized as a stigmatizing condition because it is an undesirable attribute of society s members [6]. However, the stigmatization is not restricted to children, and their relatives are stigmatized too. A good way to explain social phenomena related with lice infestation is shared vulnerability model [7]. This theoretical model uses four stages ( being ostracized, losing integrity of the self, struggling with persistence and managing strain ) to explain why some individuals have persistent infestation. Because it specifically mentions that losing integrity of the self is expressed by social isolation, modification of family relationships and feelings of guilt [7], the model can be useful as theoretical basis to understand the effect of households social networks on lice infestation. Context A population health diagnostic was conducted between 2010 and 2011 in the neighborhood of Río Escondido in Acatlipa (Morelos, Mexico) as part of an educational practicum with public health students [8]. During this diagnostics, pediculosis was identified to be a problem that mainly affected elementary school children. The adults responsible for addressing it held beliefs that were not consistent with scientific knowledge, especially in terms of the means of contagion and available treatment. Thus, it was decided to conduct a study culturally adapted to the community [9] that would serve to provide evidence of the adequate implementation of interventions to control infestation. In this context, the objective of this study was to explore the role of households social networks in the incidence of pediculosis. Methods A prospective study was conducted in which 17 school children and their social networks participated during the summer break. The baseline served as a diagnostic and ensured the treatment of infestation before follow-up. A questionnaire was administered, which included sociodemographic data and characteristics of the hair. This study was approved by the ethics commission of the Mexican National Institute of Public Health. The school children were visited in their homes during the vacation period (July and August, 2011), after receiving authorization by their guardians. Their heads were inspected, and a detailed explanation was provided to one of the parents about all the steps to be followed. The wet combing technique was used [10, 11], which was carried out with a special comb and conditioner. The hair was considered long if its length exceeded the shoulders and short if it was not >5 cm from the scalp. The hair was divided into quadrants for washing, and was taken one by one, dividing it into small locks that were combed with conditioner. As the locks were cleaned they were tied with bands. The comb was then cleaned with white paper. Following this, they were asked about close contacts during vacation, and the heads of their contacts were also inspected when they agreed to it. One of the researchers washed the hair of the child attending school and that of the members of his or her social network who had lice, with a commercial shampoo containing 5 g/100 ml permethrin. During this process, one of the parents was explained the correct dosage and application time for effective use. They were given the shampoo so as to conduct the washing three times to prevent infestation. A second inspection of the heads was conducted at school on the first day of class to evaluate infestation. Social network analysis A symmetrical matrix was plotted with the data from the social networks using the program NetDraw. [12] The density of the network was then calculated as well as the degree for each node using the program UCINET. [13] Density is a measurement used to express the number of contacts between nodes and is calculated as the ratio between contact pairs divided by the maximum number of possible contacts. [14] The degree is the number of direct contacts associated with each node [15], which for this study can be interpreted as a greater probability of P. humanus var. capitis infestation. The transmissibility index was also calculated, based on the degree and the infestation status of the nodes involved, as follows: Transmissibility Index ¼ ½# infested nodes 1 þð#non-evaluated 0:5ÞŠ= degree: Note that a score of 0.5 was assigned to nonevaluated nodes owing to uncertainty, in accordance with fuzzy logic [16]. Qualitative Comparative Analysis The use of a qualitative comparative analysis (QCA) was chosen because the size of the available sample was small. This method and analytical theory enables combining quantitative and qualitative approaches [17]. In recent years, it has begun to be used in public health studies [18, 19] to explore whether one or several combinations of attributes (configurations) 414 Journal of Tropical Pediatrics Vol. 59, No. 5
TABLE 1 Characteristics of children participating in the study. Acatlipa, Morelos, Mexico, 2011 Case Sex Hair Basal examination Degree a Infestation index b Final examination 1 Male Short No 6 0.583 No 2 Male Short No 14 0.429 No 3 Female Medium No 9 0.444 No 4 Male Short No 6 0.250 No 5 Female Long No 6 0.250 No 6 Male Short No 9 0.500 No 7 Female Medium Yes 10 0.500 Yes 8 Female Long No 2 0.750 No 9 Male Short No 4 0.250 No 10 Male Short No 9 0.500 No 11 Male Short No 9 0.278 No 12 Female Long Yes 7 0.571 No 13 Female Long No 8 0.438 Yes 14 Female Long Yes 8 0.688 Yes 15 Male Short No 2 0.250 No 16 Male Short No 8 0.500 No 17 Female Long No 5 0.600 No Set F H I D N O In bold letters the attribute with score ¼ 0.5 a Obtained during social network analysis. b Based on the number of infected nodes/degree (description in the text). FIG. 1. Social networks of children during school holidays. Square: woman; circle: man; black: pediculosis; white: no pediculosis; gray: unknown. are associated with one outcome [19, 20]. In addition to exploring several causal routes, it is also possible to include outliers, unlike conventional statistical methods [20]. The causal logic of the QCA is similar to classical epidemiological proposals including concepts as necessary and sufficient causes [21 22]. During calibration, the original variables were transformed to obtain values between 0 (total nonmembership, indicated with a lower case letter) and 1 (total membership, indicated with a capital letter). The dichotomous variables assumed a reference value in which female ¼ 1 (F), long hair ¼ 1 (H), previous infestation ¼ 1 (I) and new infestation ¼ 1 (O). Journal of Tropical Pediatrics Vol. 59, No. 5 415
TABLE 2 Truth table for configurations observed among children participating in the study a Configurations n % F*I*D*N*H 1 8.33 F*I*d*N*H 1 8.33 F*i*D*n*H 1 8.33 F*i*D*n*h 2 16.67 F*i*d*N*H 2 16.67 F*i*d*N*h 1 8.33 F*i*d*n*H 1 8.33 F*i*d*n*h 3 25.00 a Five cases were excluded because of an attribute having a score ¼ 0.5. In terms of hair, there were some cases of medium length, which were assigned a score of 0.5. Degree (D) was transformed using the standardization method and the infestation index (N) was constant because it was a ratio. The possible configurations underwent two tests. The first evaluated consistency versus inconsistency, considered to be met when the consistency of the evaluated configuration (y) is greater than that of the complementary configuration (1 y). The second test evaluated whether the consistency of each configuration exceeded a fixed value of 0.75. Then, the configurations that met both tests were determined and, finally, the most parsimonious configuration was identified. The analyses were conducted with the fuzzy macro [23] using the Stata 11 statistical program (Stata Corporation, College Station, TX, USA). Results Some of the characteristics of the children participating in the study are in Table 1. There were slightly more males (9 of 17), 9/17 had short hair, 2/17 had medium length hair and 6/17 had long hair. Note that there were three cases with pediculosis before and after the intervention (17.6%, 95% CI: 3.8 43.4), two of which presented infestation at both times. The egocentric social networks during vacation included 122 nodes, with a density of 0.0145 and a standard deviation of 0.1206 (Fig. 1). The degree presented values between 2 and 14 (median: 8). It was possible to directly evaluate 40.98% (n ¼ 50) of the total of nodes, observing 43 cases. Thus, the observed prevalence was 40.28%, which decreased to 35.25% when considering all the nodes in the network. The infestation index had values between 0.25 and 0.75 (median: 0.5), indicating the potential presence of pediculosis in all the social networks. The truth table with the eight configurations present among children shows that those most frequently observed were F*i*d*n*h, F*i*D*n*h and F*i*d*N*H (Table 2). The configurations that met both tests were F*I*d*n*H and F*I*d*N*H, which could be reduced to a single parsimonious configuration, F*I*d*H. This attained a total coverage of 0.344 and a consistency of 0.941. Discussion This study was able to identify a profile associated with the occurrence of pediculosis in the home, characterized by having had prior infestation, being female, having long hair and a low number of social network contacts. Among different populations and cultures, studies have reported consistently that pediculosis is more frequent among those who have already been infected [24, 25], who are female [2, 26, 27] or have long hair [24 26]. This can be associated with females tending to play in the home with small groups with whom they have close contact, unlike males [28]. In similar contexts, high number of women are in contact with children with lice infestation (82.35%), and mothers are the most frequent infested relative (32.5%) [29]. Moreover, a greater occurrence has also been reported after school breaks [24 26]. Nevertheless, to our knowledge, this is the first study using social network analysis (SNA) to explore the role of relationships between individuals in the transmission of lice. This could only be demonstrated by including in the analysis the simplest centrality indicator because conventional analyses only take into account those who directly participate, without considering other individuals who may have an effect [30, 31]. The finding that infested children tend to have fewer contacts in their social networks can be understood with the shared vulnerability model [7]. In this model, children with lice infestation and their families have social isolation; thus, effective interventions should include medical treatments and specific actions to facilitate the contact with wider social networks and diminish stigmatization. This study also demonstrated that insecticides are only partially effective and do not have a long-term effect if structural determinants are not modified [25]. Studies have shown that topical treatments using neurotoxic insecticides have lost a considerable degree of effectiveness [32 34]. Resistance to these insecticides can decrease their effectiveness [35, 36]. Intervention with shampoo was conducted with all school children and some members of their social networks. It is therefore possible that the lice remained in individuals from networks without the intervention, there may not have been total adherence to the intervention as indicated by the researchers or the parasites may even have been resistant to the insecticides. These situations are common and strengthen the idea that controlling lice requires enduring actions involving social determinants. This study confirms that the use of insecticides is really effective only if it is possible to assure that many 416 Journal of Tropical Pediatrics Vol. 59, No. 5
school children and their networks of contacts are treated at one time, as performed in developed countries [37]. Unfortunately, these massive experiences are expensive for developing countries, where the wealth is concentrated in small population groups. The results of the study should be interpreted considering some limitations. A small sample participated in the study, making it impossible to observe effects of the variables using conventional statistical tests. Nevertheless, the analysis enabled exploring whether there were one or more causal routes using logic based on an analysis of cases and not of their variables. This type of analysis turns out to be more coherent, using theoretical approaches such as shared vulnerability [7]. A limitation associated with QCA is that it does not facilitate the exploration of many attributes. Therefore, from the perspective of conventional statistics, the study could be considered confounding, while based on QCA logic it can be considered adequate for observing the effect of social networks. In conclusion, this study provides evidence indicating the home as a context that facilitates the transmission of pediculosis when a series of attributes are combined. SNA was a method to provide contributions to the study of conditions in which there is direct person-to-person infestation or individuals are grouped and relate according to certain shared characteristics [31]. This knowledge, created with the same community participating in a future intervention, enables generating reliable evidence that can be used later by interventions that are more specific and culturally acceptable to the same population. References 1. Feldmeier H, Heukelbach J. Epidermal parasitic skin diseases: a neglected category of poverty-associated plagues. Bull World Health Organ 2009;87:152 9. 2. Falagas ME, Matthaiou DK, Rafailidis PI, et al. Worldwide prevalence of head lice. Emerg Infec Dis 2008;14:1493 4. 3. Manrique P, Pavı a N, Rodrı guez-buenfil JC, et al. Prevalence of pediculosis capitis in children from a rural school in Yucatan, Me xico. Rev Inst Med Trop Sa o Paulo 2011;53:325 7. 4. Andresen K, McCarthy AM. A policy change strategy for head lice management. J School Nurs 2009;25: 407 16. 5. Ro zsa L, Apari P. Why infest the loved ones inherent human behavior indicates former mutualism with head lice. Parasitology 2012;139:696 700. 6. Goffman E. Stigma: Notes on the management of a spoiled identity. Englewood Cliffs, NJ: Prentice Hall, 1963. 7. Gordon SC. Shared vulnerability: a theory of caring for children with persistent head lice. J School Nurs 2007; 23:283 92. 8. Arenas-Monreal L, Cortez-Lugo M, Parada-Toro IM. Community-based participatory research and the Escuela de Salud Pu blica in Mexico. Public Health Rep 2011;126:436 40. 9. Trickett EJ, Beehler S, Deutsch C, et al. Advancing the science of community-level interventions. Am J Public Health 2011;101:1410 9. 10. Ugbomoiko-Uade S, Speare R, Heukelbach J. Selfdiagnosis of head lice infestation in rural Nigeria as a reliable rapid assessment tool for pediculosis. Open Dermatol J 2008;2:95 7. 11. Jahnke C, Bauer E, Hengge UR, Feldmeier H. Accuracy of diagnosis of pediculosis capitis: visual inspection vs wet combing. Arch Dermatol 2009;145: 309 13. 12. Borgatti SP. NetDraw: Graph Visualization Software. Harvard: Analytic Technologies, 2002. 13. Borgatti SP, Everett MG, Freeman LC. Ucinet for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies, 2002. 14. Friedkin NE. Structural cohesion and equivalence explanations of social homogeneity. Sociological Methods Research 1984;12:235 261. 15. Freeman LC. Centrality in social networks. Conceptual clarification. Soc Netw 1978/1979;1:215 39. 16. Vineis P. Methodological insights: fuzzy sets in medicine. J Epidemiol Community Health 2008;62:273 8. 17. Ragin CC. The comparative method: moving beyond qualitative and quantitative. Berkeley: University of California Press, 1987. 18. Schensul JJ, Chandran D, Singh SK, et al. The use of qualitative comparative analysis for critical event research in Alcohol and HIV in Mumbai, India. AIDS Behav 2010;14(Suppl 1):S113 25. 19. Eng S, Woodside AG. Configural analysis of the drinking man: fuzzy-set qualitative comparative analyses. Addict Behav 2012;37:541 3. 20. Rihoux B. Qualitative comparative analysis (QCA) and related systematic comparative methods. Int Sociol 2006;21:679 706. 21. Susser M. What is a cause and how do we know one? A grammar for pragmatic epidemiology. Am J Epidemiol 1991;133:635 48. 22. Idrovo AJ, Albavera-Herna ndez C, Rodrı guez- Herna ndez JM. Social epidemiology of a large outbreak of chickenpox in the Colombian sugar cane producer region: a set theory-based analysis. Cad Saude Publica 2011;27:1393 402. 23. Longest KC, Vaisey S. fuzzy: A program for performing qualitative comparative analyses (QCA) in Stata. STATA J 2008;8:79 104. 24. Rı os SM, Ferna ndez JA, Rivas F, et al. Prevalencia y factores asociados a la pediculosis en nin os de un jardı n infantil de Bogotá. Biomedica 2008;28:245 51. 25. Wolf RM, Davidovici B. Treatment of scabies and pediculosis: Facts and controversies. Clin Dermatol 2010;28:511 8. 26. Bosely HA, El-Alfy NM. Head lice infestations (Anoplura: Pediculidae) in Saudi and non-saudi school-aged children. J Egypt Soc Parasitol 2011;41: 131 40. 27. Leung AK, Fong JH, Pinto-Rojas A. Pediculosis capitis. J Pediatr Health Care 2005;19:369 73. 28. Heukelbach J, Wilcke T, Winter B, et al. Epidemiology and morbidity of scabies and pediculosis capitis in resource-poor communities in Brazil. Br J Dermatol 2005;153:150 6. 29. Casstex L, Suárez S, De la Cruz AM. Presencia de pediculosis en convivientes con nin os positivos a Journal of Tropical Pediatrics Vol. 59, No. 5 417
Pediculus capitis (Anoplura: Pediculidae). Rev Cub Med Trop 2000;52:225 7. 30. Christakis NA. Social networks and collateral health effects. Br Med J 2004;329:184 5. 31. Ma rquez-serrano M, Gonza lez-jua rez X, Castillo- Castillo LE, et al. Social network analysis to evaluate nursing interventions to improve self-care. Public Health Nurs 2012;29:361 9. 32. Mumcuoglu KY, Miller J, Galun R. Susceptibility of the human head lice and body louse, Pediculus humanus to insecticides. Insect Sci Appl 1990;11:223 6. 33. Gonza lez P, Vassena C, Zaerba E, Picollo M. Effectiveness of lotions based on essential oils from aromatic plants against permethrin resistant Pediculus humanus capitis. Arch Dermatol Res 2007;299:389 92. 34. Burguess IF. Current treatments for pediculosis capitis. Curr Opin Infect Dis 2009;22:131 6. 35. Burkhart CG, Burkhart CN. Safety and efficacy of pediculicides for head lice. Exp Opin Drug Saf 2006;5: 169 79. 36. Durand R, Bouvresse S, Berdjane Z, et al. Insecticide resistance in head lice: clinical, parasitological and genetic aspects. Clin Microbiol Infect 2012;18:338 44. 37. Rukke BA, Birkemoe T, Soleng A, et al. Head lice in Norwegian households: actions taken, costs and knowledge. PLoS One 2012;7:e32686. 418 Journal of Tropical Pediatrics Vol. 59, No. 5