Organizational culture on community health outcomes after the 2009 H1N1 pandemic.

Author:Brannen, Donald E.


Public health preparedness and response has been nominated as one of the ten great public health achievements of the first decade of the 21st century (CDC, 2011b). The 2009-2010 U.S. public health vaccination campaign has been presented in support of the nomination due to the 10,400 hospitalizations and 520 deaths that were averted by the campaign (CDC, 2011b; CDC, 2011c; Ohio Department of Health; 2011).

Four years after September 11, 2001, bioterrorism preparedness remained a high priority (Katz, Staiti, & McKenzie, 2006). From 2001 through 2005 the focus was on improving public health capacity and the second half of the decade focused on laboratory, epidemiology, surveillance, mass vaccination, prophylaxis, treatment distribution and administrative capabilities of the public health system (CDC, 2011b; Brannen & Stanley, 2004). Through flexible federal funding of bioterrorism preparedness, communities strengthened their ability to respond to public health emergencies (Cohen, Gould, & Sidel, 2004; Katz et al, 2006). Collaborative relationships developed for bioterrorism preparedness proved useful in addressing other threats, such as natural disasters and infectious disease outbreaks. However, ongoing challenges including funding constraints, inadequate surge capacity, and public health workforce shortages still persist.

Bioterrorism preparedness funding, crisis experience, and leadership have been noted as the most important determinants of local public health agencies' preparedness activities. (7) The Centers for Disease Control and Prevention (CDC) has played a mediating role by building capacity through funding key organizational leadership positions (Avery & Zabriskie-Timmerman, 2009; Katz et al, 2006). Public health has a defined leadership role at the local level in health and medical emergency support functions (CDC, 2011a; U.S. Department of Homeland Security, 2011). The key to effective public health leadership goes beyond managing within the organization. It must also invite shaping the organization's culture towards a health service of sustainability and change (Gray, 2009). Organizational performance varies with aspects of organizational culture and mixes of culture types (Reay, 2010; Singer et al, 2009). Public health related organizations can be guided through a successful cultural transition to enhance client safety, organizational adaptability, performance, and efficiency through the advocacy of their leadership (Singer et al, 2009; Stone, Bryant, & Barbarotta, 2009; Griffith, Yonas, Mason, & Havens 2010; Meagher-Stewart et al, 2010; Reay, 2010; Summerill, Pollard, & Smith, 2010;).

Given that organizational culture can be changed, the initial question should be: What cultural aspects are crucial to success during a public health emergency? To capture the essence of organizational performance; the focus should be on assessment of community level measures, establishment of targets, and utilization of available data to indicate areas directly linked to actions within the community where opportunities exist to improve the population's health (Gunasekaran, Patel, & Mcgaughey, 2004; Braz, Scavarda, & Martins, 2011; Erwin, Greene, Mays, Ricketts, & Davis, 2011). The 2009 H1N1 vaccination campaign presents an opportunity to explore the association of organizational culture aspects' to the health outcomes of the 2009 H1N1 post pandemic effect.


The study design was a prospective and linked health outcomes study of organizations and their service areas as the unit of analysis. The sampling method was a two stage methodology with the first stage comprised of a judgment sampling of emergency exercises with the second stage comprised of a voluntary simple random sample of all participating organizations. All influenza related hospitalizations were enumerated during the 2010-2011 flu season. Culture in this study refers to the organization's members' fundamental ideology and orientation including the beliefs and values shared by all members. Thus, culture is reflected in the day to day operations of the organization (Gray, 2009; Singer et al, 2009). A questionnaire was provided to evaluators and participants for selected functional and full scale exercises in West Central Ohio occurring from 2003 to 2011. The Organizational Culture Questionnaire (OCQ) uses 14 items to ask about 10 aspects of organizational culture based upon published definitions (Robbins, 1994). The questionnaire measures perceptions of individual initiative (responsibility, freedom), risk tolerance (assertive, innovative, risk), direction (objectives, performance), integrated (coordinated), support, control (oversight), identity, reward, conflict addressed, and open communication. The 'no, maybe, yes, not relevant, unknown' responses were numerically coded to 0, 0.5, 1, and not relevant and unknown were set to missing. The OCQ reliability was evaluated using Cronbach's alpha. For the reliability analysis responses were standardized with no responses set to zero; and positive responses set to 1. The inter class reliability was tested by transposing the 14 items by rater. The reliability of each rater within the event type (chemical or biological terrorism) and organizational type (government support, law enforcement, medical, public health, public safety) was then tested using a two way mixed rather than random model as only persons specialized in the rated organization's respective field were asked to be raters. The supposition is that the raters will have similar patterns of scores so the responses were checked for consistency rather than absolute agreement.

The exercise events were conducted as part of the jurisdictions' preparedness efforts either under the CDC's Bioterrorism Cooperative Agreement, the Department of Justice Metropolitan Medical Response System, or later, under the Department of Homeland Security. This research was conducted by the Greene County Combined Health District in partnership with a variety of local multi-agency and multijurisdictional response entities. The exercise events' scenario included anthrax, chemical terrorism, regional bioterrorism, or smallpox. The OCQ was a small part of the overall objectives of the exercises and was administered after the exercise either during the post exercise debriefs or later online for those unable to attend debriefs. Five cases were dropped from the pool of eligible organizations due to missing data. The final total dataset was 158 participants representing 84 organizations. Organizations were chosen to participate based on their expected roles during a public health emergency.

Statistical methods included independent t tests, linear regression, factor analysis, and multidimensional scaling. The analysis of the organizations' cultural characteristics affecting the outcome to the 2009 H1N1 pandemic was conducted using linear regression. Based upon the expected duration of efficacy of the vaccine to extend to at least the next flu season and that the vaccine component included the circulating strains within the study area, it was assumed that a communities' effectiveness during the 2009 H1N1 campaign can be measured by the amount of hospitalized influenza cases during the following flu season within the community (see Figure 1: Organizational subtypes and health outcomes).


Prior to conducting modeling, the predictor variables measuring cultural aspects were checked for intercorrelations using multicollinearity diagnostics. Factor Analysis was used to determine which variables to drop from analysis. The list of survey questions are shown in Table 1. After the intercorrelated variables were dropped, the remaining variables with the greatest p value were dropped without reintroduction in a stepwise manner until the regression model was either a single predictor or reached significance using alpha of .05. The cultural aspects were limited to those that were not intercorrelated. The Eigen values which could indicate a possible concern with intercorrelations was reduced considerably before proceeding with the final model by dropping out variables that were explained adequately by other variables within the model.

The tolerance values for the last model were more than adequate with 14 to 66%. Factor analysis revealed clear demarcations with no interlacing of correlations between components between the remaining variables.

Missing variables were handled in a list-wise manner. The dependent variable was the influenza hospitalization rate per 10 million adjusted by local population density during the 2010 to 2011 influenza season. Population density was defined as the number of persons per square mile. The population density divisor provided for greater normality of the dependent variable. Multidimensional...

To continue reading