Factors related to the development of health-promoting community activities in Spanish primary healthcare: two case–control studies

Objective Spanish primary healthcare teams have the responsibility of performing health-promoting community activities (CAs), although such activities are not widespread. Our aim was to identify the factors related to participation in those activities. Design Two case–control studies. Setting Performed in primary care of five Spanish regions. Subjects In the first study, cases were teams that performed health-promoting CAs and controls were those that did not. In the second study (on case teams from the first study), cases were professionals who developed these activities and controls were those who did not. Main outcome measures Team, professional and community characteristics collected through questionnaires (team managers/professionals) and from secondary sources. Results The first study examined 203 teams (103 cases, 100 controls). Adjusted factors associated with performing CAs were percentage of nurses (OR 1.07, 95% CI 1.01 to 1.14), community socioeconomic status (higher vs lower OR 2.16, 95% CI 1.18 to 3.95) and performing undergraduate training (OR 0.44, 95% CI 0.21 to 0.93). In the second study, 597 professionals responded (254 cases, 343 controls). Adjusted factors were professional classification (physicians do fewer activities than nurses and social workers do more), training in CAs (OR 1.9, 95% CI 1.2 to 3.1), team support (OR 2.9, 95% CI 1.5 to 5.7), seniority (OR 1.06, 95% CI 1.03 to 1.09), nursing tutor (OR 2.0, 95% CI 1.1 to 3.5), motivation (OR 3.7, 95% CI 1.8 to 7.5), collaboration with non-governmental organisations (OR 1.9, 95% CI 1.2 to 3.1) and participation in neighbourhood activities (OR 3.1, 95% CI 1.9 to 5.1). Conclusions Professional personal characteristics, such as social sensitivity, profession, to feel team support or motivation, have influence in performing health-promoting CAs. In contrast to the opinion expressed by many professionals, workload is not related to performance of health-promoting CAs.

Keywords: Health promotion, Primary health care, Health services research, Community health, Health team, health professionals

Strengths and limitations of this study
-This study examined a topic relatively unexplored, as community health promotion by primary health care.
-A case-control approach was chosen, where cases were subjects (professionals or teams) who were developing community activities, and controls those who not.
-Design allows to explore many distinct factors simultaneously, from different sources, and in two levels: Health team and professional.
-An operative definition of community activities was used, which recognizes the heterogeneity of health promotion interventions developed with the community in Spanish primary health care.
-Because of its observational design, there are limitations to establishing causal inferences.

Introduction
The goal of the Ottawa Charter for Health Promotion is to enable people to increase control over and improve their health. One way to do this is by community activities such as organizing walking groups in a neighbourhood or holding meetings with local authorities or associations to develop specific health programs. Experts at numerous international conferences have agreed on the importance of health-promoting community activities for improving public health [1][2][3] , and there is some evidence that these activities are effective 4 .
Since the implementation of health care reform in Spain during the 1980's [5][6] , primary health care teams have had responsibility for carrying out health promotion activities. These teams, located in health care centres throughout the country, are composed of physicians, community nurses, paediatricians, midwives, social workers, and other health care professionals (gynaecologists, psychologists, and psychiatrists) in a system that offers almost-universal and free health care. These conditions are ideal for an interdisciplinary approach to encourage health promotion in the community. However, as in other countries 7 , implementation of these interventions differs greatly among the different health care teams [8][9][10] , and relatively few professionals are involved [10][11][12] .
There has been limited study of factors related to the development of healthpromoting interventions in the community. These factors may be broadly classified as related to the community, to the team, or to the individual professionals. With regard to community factors, some studies indicated that the rural/urban environment 13 , size of the municipality 14 , level of social interaction within the community [14][15][16] , and presence of a focal point for participation in community health initiatives [14][15] influence the practice of community activities. The health team factors include job satisfaction 17 , management of time and setting within the center 13 , composition and organization of the team [14][15] , duration of operation 14 , intra-team support 16 , and work burden 7,18 . Studies of professionals have highlighted sex [18][19][20] , age 14,19 , professional status 13 , specific training 21-24 , motivation [16][17] , and the model applied (biomedical or psychosocial) 17 . The attitudes and beliefs of the professionals, such as trust in the effectiveness of preventive actions 17,21,24,25 , self-reported efficacy in carrying out activities 7,[21][22] , or to support community participation 14,20 are also important. An additional element is professionals' social sensitivity, as evaluated through their personal political leanings 27 . According to professionals themselves, the most important factor is the lack of time available to assume the care burden 18,21,[23][24][25][26] . All these features have been observed separately, mostly in descriptive studies.
This study presents some of the results of the Factors Related to Health-Promoting Community Activity Development in Primary Care (frAC) Project, whose complete methodology and aims were described elsewhere 28 . The purpose of this study is to identify team and community factors related to implementation of community activities in primary care and factors that may explain why only some professionals within a team participate in community activities.

Methods
Two case-control studies were performed. In the first, the cases were the teams that performed community activities and the controls were those that did not. The second study examined teams that implemented community activities to adjust for community/team factors; in this second analysis the cases were professionals who developed these activities and the controls were those who did not. The study was set in primary health care settings in five regions of Spain (Balearic Islands, Catalonia, Aragon, Madrid, and Navarra). Data collection was carried out in 2009-2010.  6 We telephoned all the health centres in the research areas asking if they had participated in community activities during the last year. If they did, we contacted a participant of the activity and gave him/her a questionnaire to confirm that the activity met our inclusion criteria. For the first study, health teams that participated in at least one community activity in the last year were selected as cases, and controls were those who had not participated. For the second study, we selected all professionals who participated in the planning of community activity as cases and professionals who did not participate in the planning as controls. In every health team we selected at least one control by every case.

Selection of cases and controls
A group of experts reached consensus on a conceptual definition of health promoting community activity 29 which was converted into an algorithm for use in this study 28 . In particular, this activity had to be a non-isolated activity carried out in the previous year, in which the professional participated on behalf of the health centre. Furthermore, it had to have involvement and active participation of the community, or the population had to have the capacity to influence the intervention, or it was a cross-sector activity involving collaboration with entities outside of health care (e.g. education, social services).

Variables
Information was gathered through three distinct questionnaires (available on request) that were given to the community, the team, and the professionals 28 . The data collected from the community (collected from secondary sources and from questionnaires to centre managers) were demographic composition; socio-economic level; degree of social interaction; health centre setting; geographical dispersion based on an ordinal variable used by the National Health System for resource allocation (from G1 [less dispersed] to G4 [more dispersed]); and existence of health boards (community participation agencies described in health legislation) 15 . The data collected from the team (collected from secondary sources and questionnaires to centre managers) were: percentages of different professionals (nurses, physicians, and paediatricians); year of opening; population of service area; professional/population and doctor/nurse ratios; availability of a space in the centre to carry out group activities and presence of a team member responsible for health education; collaboration with the area health board where applicable; training of resident physicians; nurses and nursing or physiotherapy students; evaluation of professional relationships between distinct professional disciplines (based on a Likert-type scale ranging from 1 [poor] to 5 [excellent]).
The data collected from the professionals (from individual self-administrated questionnaires) were: sex; age; profession; working situation; health care burden; size of quota allocated; average number and duration of consultations per day at the centre or at home; tutoring; participation in research or specific training in health promotion in the previous five years; years since graduation; years worked in primary care and as a member of the current health care team; autonomy to organize scheduling; arrangements with colleagues to cover health care tasks while carrying out community activities; support of fellow team members in conducting these activities; working atmosphere; job satisfaction; and selfassessed efficacy in conducting community work (based on a Likert-type scale ranging from 1 [poor] to 5 [excellent]).
Opinions and attitudes were measured by responses to various statements developed by the research team, using a Likert-type scale with four options ranging from "strongly agree" to "strongly disagree". The statements dealt with trust in the effectiveness of the community activities, motivation to carry out these activities, citizen participation in health care decision-making, need to strengthen agencies involved in citizen participation in health care, responsibility for primary care in the community, definition of professional role, link

Calculation of sample size
For the first study, we calculated that at least 91 teams would be needed in each group based on an OR of 2.5, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls. For the second study, we calculated that at least 222 professionals would be needed in each group based on an OR of 1.75, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls.

Data analysis and management
A descriptive analysis of all variables was carried out to compare their distributions in each group. Categorical variables are expressed as percentages and continuous variables as means with 95% confidence intervals (95% CIs) or as medians and percentiles depending on the distribution. The relationship with the main variable was assessed using the chisquared test for categorical variables and Student's t-test and the Mann-Whitney U nonparametric test for continuous variables. The strength of an association was expressed through unadjusted OR and its 95% CI.
We also performed a logistic regression analysis to calculate adjusted ORs. A saturated model was constructed using each variable whose p-value was less than 0.2 in the bivariate analysis. From this model, we tested and compared variations using the maximum The collinearity of the variables was examined and interactions tested. we established different models for the health care teams and for the professionals. All analyses were conducted using SPSS ver. 14 and Epidat ver. 3.1 software.

Ethical aspects
The study was approved by the research commissions in all areas where it was

Factors related to the health care team
We examined 203 health care teams (103 cases [50.7%] and 100 controls) in the first study. Table 1 shows the results of the initial bivariate analysis. The variables significantly related to team involvement in community activities were the socioeconomic level of the community, collaboration on training in physiotherapy or nursing, participation in a health board, percentage of nurses in the team, patient/physician ratio, and patient/nurse ratio.   11 Eleven teams (5.0%) were excluded from the logistic regression analysis due to missing data. The 3 variables remaining in the final model were percentage of nurses in the team (OR = 1.07 for every 1% increase; 95% CI = 1.01-1.14); socio-economic level of the community (high and medium-high vs. low and medium-low OR = 2.16; 95% CI = 1.18-3.95); and having undergraduate training at the health care centre (OR = 0.44; 95% CI = 0.21-0.93). Nagelkerke's R squared was 0.11.

Factors related to professionals
A total of 597 professionals responded to the questionnaire, 254 (42.4%) cases and 343 (57.5%) controls. Refused to participate 96 professionals (62 cases and 34 controls).

Discussion
Our study of the participation of health care workers in health-promoting activities in the community suggests that factors of the individual health care professional were more important than those of the health care team or the community. In particular, the specific Our assessment of variables related with health care team participation in community activities suggests that team characteristics had greater impact than community characteristics. One exception was the socio-economic level of the community. This makes sense in the context of the Spanish health care system, which attempts to offer equitable care to people according to their needs and community activities are important tools to reduce inequalities in health care 34 . It should be noted that we found no relationship between the level of social interaction in the community and the implementation of community activities, in contrast with the findings of other authors [15][16] .
Our results showing the important role of nurses in community activities are consistent with those of other authors [12][13]16,35 who highlighted the role of nurses in the development of community activities. This result is expected because nurses usually receive more training in community care. Another crucial group is social workers. Previous research reported the importance of social workers 12 , but their role in the community varies in different regions of Spain. Their presence seems fundamental for development of health-promoting community activities in regions where they have significant roles in the community. In Spain, community work is generally not provided as part of primary health care; instead, it is relegated to professional volunteers or specific programs. Thus, our results recognize the importance of work by nurses and social workers and points to the possibility that their roles could be 16 modified or clarified. We also found that health centres that train undergraduate nurses have a lower level of community activities but that professional nurse-trainers have a greater involvement in community activities. This apparent contradiction may be resolved by considering two factors. First, the variable at the health centre level included training in various disciplines (from nursing to physiotherapy) without distinction, even though there is clearly difference among these in communitarian tradition. Second, the need for increased training would make it difficult for the health team to engage in community activities although, among the cases trainers carried out more activities than others. Health care workers who worked for more years at a centre were also more likely to perform community activities.
This may be because medium-and long-term projects require continuity and stability, and it is more difficult for professionals who continually change centres or have temporary contracts to participate in community activities.
The pressures of working in health care and the heavy work burden were the main arguments offered by professionals when asked about their low participation in health promotion and prevention activities 18,21,23-26 . Our bivariate analysis supported this relationship, although the multivariate analysis did not. This implies that the professionals' perception was incorrect. In fact, the pressure of working in health care and the heavy work burden are not associated with low participation in community activities.
Our results indicated the importance of the opinions and attitudes of professionals when performing community activities, in accordance with other studies 14,[20][21]

Strengths and limitations
This study had an observational design, so we could not definitively establish causal relationships. Similarly, our definition of "community activities" may be a limitation because this concept can be defined or perceived in different ways; we made an attempt to control for this by use of a definition generated by expert consensus, which we then transformed into an algorithm. The questionnaires we used were ad hoc and some of the data regarding team variables, such as degree of social interaction in the community and the relationship between professionals, were provided by team managers. This may have led to a bias in information gathering due to a lack of knowledge or other unknown factors. Some variables that we did not examine could have explained participation in community activities, such as support from 18 primary care management, the health care administration and the centres' own coordinators or nursing managers. It would be valuable to assess the effect of these variables in future studies of the effect of team activities and priorities.
The main strengths of this study are that it examined a facet of health care that is relatively unexplored; the study subjects were from 5 different regions of Spain; the sample size was large; and it considered many distinct factors simultaneously. Furthermore, our design of the second study, in which cases and controls were matched by teams, adjusted for the effect of team and community variables, so a multilevel analysis was not necessary.

Implications for health care practice
Health care administrators who want to develop health-promotion activities for the community must provide visible and formal support for community initiatives, define better the expected communitarian tasks of each sanitarian professional, and provide job stability for all participating workers. Specific training is also important for the formation of healthpromoting community activities. It may seem that attitudes, social sensitivity, or personal opinions are not modifiable; however, in our opinion, it is possible to modify these factors during undergraduate education. So, community perspective and their associated responsibility should be included in undergraduate sanitary formation.
Keywords: Health promotion, Primary health care, Health services research, Community health, Health team, health professionals

Strengths and limitations of this study
-This study examined a topic relatively unexplored, as community health promotion by primary health care.
-A case-control approach was chosen, where cases were subjects (professionals or teams) who were developing community activities, and controls those who not.
-Design allows to explore many distinct factors simultaneously, from different sources, and in two levels: Health team and professional.
-An operative definition of community activities was used, which recognizes the heterogeneity of health promotion interventions developed with the community in Spanish primary health care.
-Because of its observational design, there are limitations to establishing causal inferences.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  implementation of these interventions differs greatly among the different health care teams [9][10][11] , and relatively few professionals are involved 11-13 . There has been limited studies of the factors related to the development of healthpromoting interventions in the community. These factors may be broadly classified as related to the community, to the team, or to the individual professionals. With regard to community factors, some studies indicated that the rural/urban environment 14 , size of the municipality 15 , level of social interaction within the community [15][16][17] , and presence of a focal point for participation in community health initiatives [15][16] influence the practice of health  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  This study presents some of the results of the Factors Related to Health-Promoting Community Activity Development in Primary Care (frAC) Project, whose complete methodology and aims were described elsewhere 29 . The purpose of this study is to identify primary health care team and community factors related to implementation of health promoting community activities in primary care and factors that may explain why only some professionals within a team participate in those community activities.

Methods
Two case-control studies were performed. In the first, the cases were the teams that performed health promoting community activities and the controls were those that did not.
The second study examined only teams that implemented these community activities to adjust for community/team factors; in this second analysis the cases were professionals who developed these activities and the controls were those who did not. The study was set in

Selection of cases and controls
We used a conceptual definition of health promoting community activity reached in a previous study 30-31 . This was converted into an algorithm for use in this study 29 : this activity had to be a non-isolated activity carried out in the previous year, in which the professional participated on behalf of the health centre. Furthermore, it had to have involvement and active participation of the community, or the population had to have the capacity to influence the intervention, or it was a cross-sector activity involving collaboration with entities outside of health care (e.g. education, social services). Health preventive interventions which does not have a health promotion approach were excluded.
We telephoned all the health centres in the research areas asking if they had participated in health promoting community activities during the last year. If they did, we contacted a participant of the activity and gave him/her a questionnaire to confirm that the activity met our inclusion criteria. For the first study, health teams that participated in at least one community activity in the last year were selected as cases, and controls were those who had not participated. For the second study, we selected all professionals who participated in the implementation of those community activities as cases and professionals who did not participate as controls. In every health team we selected at least one control for every case.

Variables
Information was gathered through three distinct questionnaires (available on request) that were given to the community, the team, and the professionals 29 . Description of health promoting community activities are described elsewhere 30 .
The data collected from the community (from secondary sources and from questionnaires to centre managers) were: demographic composition; socio-economic level; degree of social

Calculation of sample size
For the first study, we calculated that at least 91 teams would be needed in each group based on an OR of 2.5, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls. For the second study, we calculated that at least 222 professionals would be needed in each group based on an OR of 1.75, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls. We also performed a logistic regression analysis to calculate adjusted ORs. We constructed a first model using each variable whose p-value was less than 0.2 in the bivariate analysis. From this model, we tested and compared variations using the maximum likelihood method, eliminating variables that had p-values less than 0.05 in the Wald test. The collinearity of the variables was examined and interactions tested. We established a different model for the health care teams and for the professionals. All analyses were conducted using SPSS ver. 14 and Epidat ver. 3.1 software.

Factors related to the health care team
We examined 203 health care teams (103 cases [50.7%] and 100 controls) in the first study. Table 1 shows the results of the initial bivariate analysis. The variables significantly related to team involvement in community activities were the socioeconomic level of the community, collaboration on training in physiotherapy or nursing, participation in a health board, percentage of nurses in the team, patient/physician ratio, and patient/nurse ratio.  Tables 2 and 3 show the results of the bivariate descriptive analysis. Table 2 shows that the variables related to involvement in health promoting community activities were sex, profession, training, self-confidence, and work situation.
Also years of experience, age, workload, practice style and political orientation seems to be related. Table 3 summarizes the responses of the professionals' opinions and attitudes to health promotion, community participation, and social sensitivity.

Discussion
Our study of the participation of health care workers in health-promoting activities in the community suggests that factors of the individual health care professional were more  care; instead, it is relegated to professional volunteers or specific programs. Thus, our results recognize the importance of work by nurses and social workers and points to the possibility that their roles could be modified or clarified. We also found that health centres that train undergraduate nurses have a lower level of community activities but that professional nurse- Second, the need for increased training would make it difficult for the health team to engage in community activities although, among the cases trainers carried out more activities than others. Health care workers who worked for more years at a centre were also more likely to perform community activities. This may be because medium-and long-term projects require continuity and stability, and it is more difficult for professionals who continually change centres or have temporary contracts to participate in health promoting community activities.
The pressures of working in health care and the heavy work burden were the main arguments offered by professionals when asked about their low participation in health promotion and prevention activities 19,22,[24][25][26][27] . Our bivariate analysis supported this relationship, although the multivariate analysis did not. This implies that the professionals' perception was incorrect. In fact, the pressure of working in health care and the heavy work burden are not associated with low participation in health promoting community activities.
Our results indicated the importance of the opinions and attitudes of professionals when performing health promoting community activities, in accordance with other studies 15,[21][22] . In fact, all variables regarding the opinions and attitudes of health care workers were significantly different between cases and controls. This suggests that individuals who perform community activities have similar attitudes: they employ a biopsychosocial 12,37 practice model, are more oriented towards health promotion, show recognition of the need for citizen participation in health services, and are more left-leaning in ideology. We also found that social sensitivity was an important variable in the model for professionals. This variable measures involvement of the professional with the community beyond consultation work. It is We found that having support within the health care team is associated with performing health promoting community activities. This finding should be highlighted, because all the professionals in this study were selected from the centres that engage in community activities so that, presumably, all cases can count on some support from their centres. It may be that some professionals were unaware of what their colleagues were doing, or that there were barriers to working with particular colleagues. Regardless of the explanation, this result suggests that application of the teamwork model 38 , an assumed key feature of primary health care, has certain weaknesses.

Strengths and limitations
This study had an observational design, so we could not definitively establish causal relationships. Similarly, our definition of "health promoting community activities" may be a limitation because this concept can be defined or perceived in different ways; we made an attempt to control for this by use of a definition generated by expert consensus, which we then transformed into an algorithm. The questionnaires we used were ad hoc and some of the data regarding team variables, such as degree of social interaction in the community and the relationship between professionals, were provided by team managers. This may have led to a bias in information gathering due to a lack of knowledge or other unknown factors. Some variables that we did not examine could have explained participation in community activities,  19 such as support from primary care management, the health care administration and the centres' own coordinators or nursing managers. It would be valuable to assess the effect of these variables in future studies of the effect of team activities and priorities.
The main strengths of this study are that it examined a facet of health care that is relatively unexplored; the study subjects were from 5 different regions of Spain; the sample size was large; and it considered many distinct factors simultaneously. Furthermore, our design of the second study, in which cases and controls were matched by teams, adjusted for the effect of team and community variables, so a multilevel analysis was not necessary.

Implications for health care practice
Health care administrators who want to develop health-promotion activities for the community must provide visible and formal support for community initiatives, define better the expected communitarian tasks of each sanitarian professional, and provide job stability for all participating workers. Specific training is also important for the formation of healthpromoting community activities. It may seem that attitudes, social sensitivity, or personal opinions are not modifiable; however, in our opinion, it is possible to modify these factors during undergraduate education. So, community perspective and their associated responsibility should be included in undergraduate sanitary formation.

Conclusions
A key responsibility of primary health care is provision of health-promoting community activities. We found that the actual implementation of these activities depends more on the characteristics of individual professionals than on the characteristics and composition of the health care team or the community. There were also significant differences among different types of professionals, and the contributions of nurses and social workers are fundamental. The pressures perceived by members of the health care team were not associated with involvement in community activities.         -A case-control approach was chosen, where cases were subjects (professionals or teams) who were developing community activities, and controls those who not.
-Design allows to explore many distinct factors simultaneously, from different sources, and in two levels: Health team and professional.
-An operative definition of community activities was used, which recognizes the heterogeneity of health promotion interventions developed with the community in Spanish primary health care.
-Because of its observational design, there are limitations to establishing causal inferences. Since the implementation of health care reform in Spain during the 1980's 6-7 , primary health care teams have had responsibility for carrying out health promotion activities. These teams, located in health care centres throughout the country, are composed of physicians, community nurses, paediatricians, midwives, social workers, and other health care professionals (gynaecologists, psychologists, and psychiatrists) in a system that offers almost-universal and free health care. These conditions are ideal for an interdisciplinary approach to encourage health promoting activities. However, as in other countries 8 , implementation of these interventions differs greatly among the different health care teams [9][10][11] , and relatively few professionals are involved [11][12][13] .
There has been limited studies of the factors related to the development of healthpromoting interventions in the community. These factors may be broadly classified as related to the community, to the team, or to the individual professionals. With regard to community factors, some studies indicated that the rural/urban environment 14 , size of the municipality 15 , level of social interaction within the community [15][16][17] , and presence of a focal point for participation in community health initiatives [15][16] influence the practice of health This study presents some of the results of the Factors Related to Health-Promoting Community Activity Development in Primary Care (frAC) Project, whose complete methodology and aims were described elsewhere 29 . The purpose of this study is to identify primary health care team and community factors related to implementation of health promoting community activities in primary care and factors that may explain why only some professionals within a team participate in those community activities.

Methods
Two case-control studies were performed. In the first, the cases were the teams that performed health promoting community activities and the controls were those that did not.
The second study examined only teams that implemented these community activities to adjust for community/team factors; in this second analysis the cases were professionals who developed these activities and the controls were those who did not. The study was set in

Selection of cases and controls
We used a conceptual definition of health promoting community activity reached in a previous study 30-31 . This was converted into an algorithm for use in this study 29 : this activity had to be a non-isolated activity carried out in the previous year, in which the professional participated on behalf of the health centre. Furthermore, it had to have involvement and active participation of the community, or the population had to have the capacity to influence the intervention, or it was a cross-sector activity involving collaboration with entities outside of health care (e.g. education, social services). Health preventive interventions which does not have a health promotion approach were excluded.
We telephoned all the health centres in the research areas asking if they had participated in health promoting community activities during the last year. If they did, we contacted a participant of the activity and gave him/her a questionnaire to confirm that the activity met our inclusion criteria. For the first study, health teams that participated in at least one community activity in the last year were selected as cases, and controls were those who had not participated. For the second study, we selected all professionals who participated in the implementation of those community activities as cases and professionals who did not participate as controls. In every health team we selected at least one control for every case.

Variables
Information was gathered through three distinct questionnaires (available on request) that were given to the community, the team, and the professionals 29 . Description of health promoting community activities are described elsewhere 30 .
The data collected from the community (from secondary sources and from questionnaires to centre managers) were: demographic composition; socio-economic level; degree of social

Calculation of sample size
For the first study, we calculated that at least 91 teams would be needed in each group  We also performed a logistic regression analysis to calculate adjusted ORs. We constructed a first model using each variable whose p-value was less than 0.2 in the bivariate analysis. From this model, we tested and compared variations using the maximum likelihood method, eliminating variables that had p-values less than 0.05 in the Wald test. The collinearity of the variables was examined and interactions tested. We established a different model for the health care teams and for the professionals. All analyses were conducted using SPSS ver. 14 and Epidat ver. 3.1 software.

Ethical aspects
The study was approved by

Factors related to the health care team
We examined 203 health care teams (103 cases [50.7%] and 100 controls) in the first study. Table 1 shows the results of the initial bivariate analysis. The variables significantly related to team involvement in community activities were the socioeconomic level of the community, collaboration on training in physiotherapy or nursing, participation in a health board, percentage of nurses in the team, patient/physician ratio, and patient/nurse ratio.  Tables 2 and 3 show the results of the bivariate descriptive analysis. Table 2 shows that the variables related to involvement in health promoting community activities were sex, profession, training, self-confidence, and work situation.
Also years of experience, age, workload, practice style and political orientation seems to be related. Table 3 summarizes the responses of the professionals' opinions and attitudes to health promotion, community participation, and social sensitivity.

Discussion
Our study of the participation of health care workers in health-promoting activities in the community suggests that factors of the individual health care professional were more  that their roles could be modified or clarified. We also found that health centres that train undergraduate nurses have a lower level of community activities but that professional nurse- Second, the need for increased training would make it difficult for the health team to engage in community activities although, among the cases trainers carried out more activities than others. Health care workers who worked for more years at a centre were also more likely to perform community activities. This may be because medium-and long-term projects require continuity and stability, and it is more difficult for professionals who continually change centres or have temporary contracts to participate in health promoting community activities.
The pressures of working in health care and the heavy work burden were the main arguments offered by professionals when asked about their low participation in health promotion and prevention activities 19,22,24-27 . Our bivariate analysis supported this relationship, although the multivariate analysis did not. This implies that the professionals' perception was incorrect. In fact, the pressure of working in health care and the heavy work burden are not associated with low participation in health promoting community activities.
Our results indicated the importance of the opinions and attitudes of professionals when performing health promoting community activities, in accordance with other studies 15,[21][22] . In fact, all variables regarding the opinions and attitudes of health care workers were significantly different between cases and controls. This suggests that individuals who perform community activities have similar attitudes: they employ a biopsychosocial 12,37 practice model, are more oriented towards health promotion, show recognition of the need for citizen participation in health services, and are more left-leaning in ideology. We also found that social sensitivity was an important variable in the model for professionals. This variable measures involvement of the professional with the community beyond consultation work. It is

Strengths and limitations
This study had an observational design, so we could not definitively establish causal relationships. Similarly, our definition of "health promoting community activities" may be a limitation because this concept can be defined or perceived in different ways; we made an attempt to control for this by use of a definition generated by expert consensus, which we then transformed into an algorithm. The questionnaires we used were ad hoc and some of the data regarding team variables, such as degree of social interaction in the community and the relationship between professionals, were provided by team managers. This may have led to a bias in information gathering due to a lack of knowledge or other unknown factors. Some variables that we did not examine could have explained participation in community activities,  19 such as support from primary care management, the health care administration and the centres' own coordinators or nursing managers. It would be valuable to assess the effect of these variables in future studies of the effect of team activities and priorities.
The main strengths of this study are that it examined a facet of health care that is relatively unexplored; the study subjects were from 5 different regions of Spain; the sample size was large; and it considered many distinct factors simultaneously. Furthermore, our design of the second study, in which cases and controls were matched by teams, adjusted for the effect of team and community variables, so a multilevel analysis was not necessary.

Implications for health care practice
Health care administrators who want to develop health-promotion activities for the community must provide visible and formal support for community initiatives, define better the expected communitarian tasks of each sanitarian professional, and provide job stability for all participating workers. Specific training is also important for the formation of healthpromoting community activities. It may seem that attitudes, social sensitivity, or personal opinions are not modifiable; however, in our opinion, it is possible to modify these factors during undergraduate education. So, community perspective and their associated responsibility should be included in undergraduate sanitary formation.

Conclusions
A key responsibility of primary health care is provision of health-promoting community activities. We found that the actual implementation of these activities depends more on the characteristics of individual professionals and on some characteristics of health care team or community than on his professional responsibility like primary health care workers. There were also significant differences among different types of professionals, and the contributions of nurses and social workers are fundamental. The pressures perceived by

Acknowledgements
We want to express thanks to Magdalena Esteva for her patient reading of the manuscript.

Competing interests
All         -A case-control approach was chosen, where cases were subjects (professionals or teams) who were developing health promoting community activities, and controls those who did not.
-Design allows explor many distinct factors simultaneously, from different sources, and in two levels: Health team and professional.
-An operative definition of community activities was used, which recognizes the heterogeneity of health promotion interventions developed with the community in Spanish primary health care settings.
-Because of its observational design, there are limitations to establishing causal inferences. Since the implementation of health care reform in Spain during the 1980's 6-7 , primary health care teams have had responsibility for carrying out health promotion activities. These teams, located in health care centres throughout the country, are composed of physicians, community nurses, paediatricians, midwives, social workers, and other health care professionals (gynaecologists, psychologists, and psychiatrists) in a system that offers almost-universal and free health care. These conditions are ideal for an interdisciplinary approach to encourage health promoting community activities. However, as in other countries 8 , implementation of these interventions differs greatly among the different health care teams [9][10][11] , and relatively few professionals are involved [11][12][13] .
There has been limited studies of the factors related to the development of healthpromoting interventions in the community. These factors may be broadly classified as those related to the community, to the team, or to the individual professionals. With regard to community factors, some studies indicated that the rural/urban environment 14 , size of the municipality 15 , level of social interaction within the community [15][16][17] , and presence of a focal point for participation in community health initiatives [15][16] influence the practice of health This study presents some of the results of the Factors Related to Health-Promoting Community Activity Development in Primary Care (frAC) Project, whose complete methodology and aims were described elsewhere 29 . The purpose of this study is to identify primary health care team and community factors related to implementation of health promoting community activities in primary care and factors that may explain why only some professionals within a team participate in those community activities.

Methods
Two case-control studies were performed. In the first, the cases were the teams that performed health promoting community activities and the controls were those that did not.
The second study examined only teams that implemented these community activities to adjust for community/team factors; in this second analysis the cases were professionals who developed these activities and the controls were those who did not. The study was set in

Selection of cases and controls
We used a conceptual definition of health promoting community activity reached in a previous study 30-31 . This was converted into an algorithm for use in this study 29 : this activity had to be a non-isolated activity carried out in the previous year, in which the professional participated on behalf of the health centre. Furthermore, it had to have involvement and active participation of the community, or the population had to have the capacity to influence the intervention, or it was a cross-sector activity involving collaboration with entities outside of health care (e.g. education, social services). Health preventive interventions which does not have a health promotion approach were excluded.
We telephoned all the health centres in the research areas asking if they had participated in health promoting community activities during the last year. If they did, we contacted a participant of the activity and gave him/her a questionnaire to confirm that the activity met our inclusion criteria. For the first study, health teams that participated in at least one community activity in the last year were selected as cases, and controls were those who had not participated. For the second study, we selected all professionals who participated in the implementation of those community activities as cases and professionals who did not participate as controls. In every health team we selected at least one control for every case.

Variables
Information was gathered through three distinct questionnaires (available on request) that were given to the community, the team, and the professionals 29 . Description of health promoting community activities are described elsewhere 30 .
The data collected from the community (from from questionnaires to centre managers and secondary sources) were: demographic composition; socio-economic level; degree of social

Calculation of sample size
For the first study, we calculated that at least 91 teams would be needed in each group  We also performed a logistic regression analysis to calculate adjusted ORs. We constructed a first model using each variable whose p-value was less than 0.2 in the bivariate analysis. From this model, we tested and compared variations using the maximum likelihood method, eliminating variables that had p-values less than 0.05 in the Wald test. The collinearity of the variables was examined and interactions tested. We established a different model for the health care teams and for the professionals. All analyses were conducted using SPSS ver. 14 and Epidat ver. 3.1 software.

Ethical aspects
The study was approved by

Factors related to the health care team
We examined 203 health care teams (103 cases [50.7%] and 100 controls) in the first study. Table 1 shows the results of the initial bivariate analysis. The variables significantly related to team involvement in community activities were the socioeconomic level of the community, collaboration on training in physiotherapy or nursing, participation in a health board, percentage of nurses in the team, patient/physician ratio, and patient/nurse ratio.    Tables 2 and 3 show the results of the bivariate descriptive analysis. Table 2 shows that the variables related to involvement in health promoting community activities were sex, profession, training, self-confidence, and work situation.
Second, the need for increased training would make it difficult for the health team to engage in community activities although, among the cases trainers carried out more activities than others. Health care workers who worked for more years at a centre were also more likely to perform community activities. This may be because medium-and long-term projects require continuity and stability, and it is more difficult for professionals who continually change centres or have temporary contracts to participate in health promoting community activities.
The pressures of working in health care and the heavy work burden were the main arguments offered by professionals when asked about their low participation in health promotion and prevention activities 19,22,[24][25][26][27] . Our bivariate analysis supported this relationship, although the multivariate analysis did not. This implies that the professionals' perception was incorrect. In fact, the pressure of working in health care and the heavy work burden are not associated with low participation in health promoting community activities.
Our results indicated the importance of the opinions and attitudes of professionals when performing health promoting community activities, in accordance with other studies 15,[21][22] . In fact, all variables regarding the opinions and attitudes of health care workers were significantly different between cases and controls. This suggests that individuals who perform community activities have similar attitudes: they employ a biopsychosocial 12,37 practice model, are more oriented towards health promotion, show recognition of the need for citizen participation in health services, and are more left-leaning in ideology. We also found that social sensitivity was an important variable in the model for professionals. This variable measures involvement of the professional with the community beyond consultation work. It is  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y 18 not clear how some practices, responsibility for which is stipulated in primary care legislation and which should be similar for all professionals to assure fairness within the system, can be affected by a professional's personal discourse. Most health care professionals that we interviewed -both cases and controls --said that their community roles were not well defined. This suggests that greater priority should be given to develop health promoting community activities in primary health care, and to clarify the role and responsibility of every professional about this question.
We found that having support within the health care team is associated with performing health promoting community activities. This finding should be highlighted, because all the professionals in this study were selected from the centres that engage in community activities so that, presumably, all cases can count on some support from their centres. It may be that some professionals were unaware of what their colleagues were doing, or that there were barriers to working with particular colleagues. Regardless of the explanation, this result suggests that application of the teamwork model 38 , an assumed key feature of primary health care, has certain weaknesses.

Strengths and limitations
This study had an observational design, so we could not definitively establish causal relationships. Similarly, our definition of "health promoting community activities" may be a limitation because this concept can be defined or perceived in different ways; we made an attempt to control for this by use of a definition generated by expert consensus, which we then transformed into an algorithm. The questionnaires we used were ad hoc and some of the data regarding team variables, such as degree of social interaction in the community and the relationship between professionals, were provided by team managers. This may have led to a bias in information gathering due to a lack of knowledge or other unknown factors. Some variables that we did not examine could have explained participation in community activities, such as support from primary care management, the health care administration and the centres' own coordinators or nursing managers. It would be valuable to assess the effect of these variables in future studies of the effect of team activities and priorities.
The main strengths of this study are that it examined a facet of health care that is relatively unexplored; the study subjects were from 5 different regions of Spain; the sample size was large; and it considered many distinct factors simultaneously. Furthermore, our design of the second study adjusted the effect of team and community variables matching cases and controls by teams, so a multilevel analysis was not necessary.Implications for health care practice Health care administrators who want to develop health-promotion activities for the community must provide visible and formal support for community initiatives, define better the expected communitarian tasks of each sanitarian professional, and provide job stability for all participating workers. Specific training is also important for the formation of healthpromoting community activities. It may seem that attitudes, social sensitivity, or personal opinions are not modifiable; however, in our opinion, it is possible to modify these factors during undergraduate education. So, community perspective and their associated responsibility should be included in healthcare workers training
-Design allows exploration of many distinct factors simultaneously, from different sources, and in two levels: Health team and professional.
-An operative definition of community activities was used, which recognizes the heterogeneity of health promotion interventions developed with the community in Spanish primary health care settings.
-Because of its observational design, there are limitations to establishing causal inferences. Since the implementation of health care reform in Spain during the 1980's 6-7 , primary health care teams have had responsibility for carrying out health promotion activities. These teams, located in health care centres throughout the country, are composed of physicians, community nurses, paediatricians, midwives, social workers, and other health care professionals (gynaecologists, psychologists, and psychiatrists) in a system that offers almost-universal and free health care. These conditions are ideal for an interdisciplinary approach to encourage health promoting community activities. However, as in other countries 8 , implementation of these interventions differs greatly among the different health care teams [9][10][11] , and relatively few professionals are involved [11][12][13] .
There has been limited studies of the factors related to the development of healthpromoting interventions in the community. These factors may be broadly classified as those related to the community, to the team, or to the individual professionals. With regard to community factors, some studies indicated that the rural/urban environment 14 , size of the municipality 15 , level of social interaction within the community [15][16][17] , and presence of a focal point for participation in community health initiatives [15][16] influence the practice of health This study presents some of the results of the Factors Related to Health-Promoting Community Activity Development in Primary Care (frAC) Project, whose complete methodology and aims were described elsewhere 29 . The purpose of this study is to identify primary health care team and community factors related to implementation of health promoting community activities in primary care and factors that may explain why only some professionals within a team participate in those community activities.

Methods
Two case-control studies were performed. In the first, the cases were the teams that performed health promoting community activities and the controls were those that did not.
The second study examined only teams that implemented these community activities to adjust for community/team factors; in this second analysis the cases were professionals who developed these activities and the controls were those who did not. The study was set in

Selection of cases and controls
We used a conceptual definition of health promoting community activity reached in a previous study [30][31] . This was converted into an algorithm for use in this study 29 : this activity had to be a non-isolated activity carried out in the previous year, in which the professional participated on behalf of the health centre. Furthermore, it had to have involvement and active participation of the community, or the population had to have the capacity to influence the intervention, or it was a cross-sector activity involving collaboration with entities outside of health care (e.g. education, social services). Health preventive interventions which does not have a health promotion approach were excluded.
We telephoned all the health centres in the research areas asking if they had participated in health promoting community activities during the last year. If they did, we contacted a participant of the activity and gave him/her a questionnaire to confirm that the activity met our inclusion criteria. For the first study, health teams that participated in at least one community activity in the last year were selected as cases, and controls were those who had not participated. For the second study, we selected all professionals who participated in the implementation of those community activities as cases and professionals who did not participate as controls. In every health team we selected at least one control for every case.

Variables
Information was gathered through three distinct questionnaires (available on request) that were given to the community, the team, and the professionals 29 . Description of health promoting community activities are described elsewhere 30 .

Calculation of sample size
For the first study, we calculated that at least 91 teams would be needed in each group based on an OR of 2.5, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls. For the second study, we calculated that at least 222 professionals would be needed in each group based on an OR of 1.75, an alpha risk of 0.05, a beta of 0.2, and an expected proportion of 0.5 in the controls.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   9 variables. The strength of an association was expressed through unadjusted OR and its 95% CI.

Data analysis and management
We also performed a logistic regression analysis to calculate adjusted ORs. We constructed a first model using each variable whose p-value was less than 0.2 in the bivariate analysis. From this model, we tested and compared variations using the maximum likelihood method, eliminating variables that had p-values less than 0.05 in the Wald test. The collinearity of the variables was examined and interactions tested. We established a different model for the health care teams and for the professionals. All analyses were conducted using SPSS ver. 14 and Epidat ver. 3.1 software.

Ethical aspects
The study was approved by
The pressures of working in health care and the heavy work burden were the main arguments offered by professionals when asked about their low participation in health promotion and prevention activities 19,22,[24][25][26][27] . Our bivariate analysis supported this relationship, although the multivariate analysis did not. This implies that the professionals' perception was incorrect. In fact, the pressure of working in health care and the heavy work burden are not associated with low participation in health promoting community activities.
Our results indicated the importance of the opinions and attitudes of professionals when performing health promoting community activities, in accordance with other studies 15,[21][22] . In fact, all variables regarding the opinions and attitudes of health care workers were significantly different between cases and controls. This suggests that individuals who perform community activities have similar attitudes: they employ a biopsychosocial 12,37 practice model, are more oriented towards health promotion, show recognition of the need for citizen participation in health services, and are more left-leaning in ideology. We also found that social sensitivity was an important variable in the model for professionals. This variable measures involvement of the professional with the community beyond consultation work. It is and which should be similar for all professionals to assure fairness within the system, can be affected by a professional's personal discourse. Most health care professionals that we interviewed -both cases and controls --said that their community roles were not well defined. This suggests that greater priority should be given to develop health promoting community activities in primary health care, and to clarify the role and responsibility of every professional about this question.
We found that having support within the health care team is associated with performing health promoting community activities. This finding should be highlighted, because all the professionals in this study were selected from the centres that engage in community activities so that, presumably, all cases can count on some support from their centres. It may be that some professionals were unaware of what their colleagues were doing, or that there were barriers to working with particular colleagues. Regardless of the explanation, this result suggests that application of the teamwork model 38 , an assumed key feature of primary health care, has certain weaknesses.

Strengths and limitations
This study had an observational design, so we could not definitively establish causal relationships. Similarly, our definition of "health promoting community activities" may be a limitation because this concept can be defined or perceived in different ways; we made an attempt to control for this by use of a definition generated by expert consensus, which we then transformed into an algorithm. The questionnaires we used were ad hoc and some of the data regarding team variables, such as degree of social interaction in the community and the relationship between professionals, were provided by team managers. This may have led to a bias in information gathering due to a lack of knowledge or other unknown factors. Some variables that we did not examine could have explained participation in community activities,  19 such as support from primary care management, the health care administration and the centres' own coordinators or nursing managers. It would be valuable to assess the effect of these variables in future studies of the effect of team activities and priorities.
The main strengths of this study are that it examined a facet of health care that is relatively unexplored; the study subjects were from 5 different regions of Spain; the sample size was large; and it considered many distinct factors simultaneously. Furthermore, our design of the second study adjusted the effect of team and community variables matching cases and controls by teams, so a multilevel analysis was not necessary. Given the extension of the achieved sample size and the inclusion of regions with different size, our results could be generalized to Spanish primary health care.

Implications for health care practice
Health care administrators who want to develop health-promotion activities for the community must provide visible and formal support for community initiatives, define better the expected communitarian tasks of each health professional, and provide job stability for all participating workers. Specific training is also important for the formation of healthpromoting community activities. It may seem that attitudes, social sensitivity, or personal opinions are not modifiable; however, in our opinion, it is possible to modify these factors during undergraduate education. So, community perspective and their associated responsibility should be included in healthcare workers training