We conducted a cross-sectional study of factors associated with health-related quality of life in a representative sample of military policemen from Salvador, Bahia, Brazil. The accessible population was composed of all the 3,500 military police officers from 27 battalions. The study included only males, engaged in visible patrolling in public spaces in the city, and excluded those who were on medical leave, engaged in administrative functions, or in specific situations not related to their core-activity. This study is part of a larger study that evaluated other health aspects in military police officers from Salvador .
Between February and April 2014, a trained researcher applied an individual, structured questionnaire to the policemen during their working hours. Information was collected about age, education, marital status, housing, car ownership, police rank, working day, alcohol consumption, smoking, frequency of vigorous physical activity and weight and height, in order to calculate body mass index. Obesity was defined as body mass index ≥30.0 . The policeman’s own perception of their work ability was measured by the Work Ability Index - WAI . The Work Ability Index questionnaire has been translated  and validated for the Brazilian population, with satisfactory construct validity and reliability  and test-retest reliability . WAI has seven dimensions: current work ability compared to lifetime best (0-10 points), work ability in relation to demands of the job (2-20 points), number of diseases diagnosed by a physician (1-7 points), estimated work impairment due to diseases (1-6 points), sick leave over the past 12 months (1-5 points), own prognosis of work ability 2 years from now (1-7 points), and mental resources (1-4 points). The WAI index is calculated through the sum of points over the seven dimensions, varying from 7 to 49 points. The score can be classified according to four work ability categories: poor (7-27 points), moderate (28-36 points), good (37-43 points), and excellent (44-49 points) . For the purposes of this study, poor work ability was codified as Yes (WAI 7-27 points) or No (WAI 28-49 points). Health-related quality of life was evaluated using the 36-Item Short Form Survey (SF-36) .
We selected a random sample, proportionally stratified by the 27 battalions. As recommended by the SF-36 manual , we used Cohen's formula  to calculate a sample size capable of detecting differences of 5.0 points between the sample mean and the mean of a fixed standard. To estimate the population’s standard deviations, we adopted the values for males aged 25-34, 35-44, and 44-54 years old, as demonstrated in Table 10.3 of the SF-36 manual . We used the value 30.24, related to the Role Physical domain, for men aged 44-54 years, since this was the largest of the three age groups and the eight SF-36 domains. The estimated minimal sample size was n=289, but this was deliberately inflated by 20%, resulting in a desired sample size of n=347 policemen. There were no refusals, however 18 individuals returned questionnaires with incomplete answers and were excluded from the study. The final sample size was composed of 329 policemen.
Data processing and analysis
Based on answers to the SF-36 questionnaire, eight multi-item scales (physical functioning, role limitations due to physical problems, bodily pain, general health perceptions, vitality, social functioning, role limitations due to emotional problems and mental health) were constructed and subsequently aggregated into two Component Summaries: the Physical Component Summary and the Mental Component Summary. The eight scales were scored using raw data (0 to 100 algorithms), while the respective normalized scores and the scores for the two summary measures was performed using QualityMetric Health OutcomesTM Scoring Software 4.0 . The study was licensed by QualityMetric Health OutcomesTM under number QM025904. The normalization procedure transforms raw scores into a mean of 50 and a standard deviation of 10, taking the United States of America general population as a reference. This transformation achieves the same mean and standard deviation for all SF-36 scales and summary measures. Because of their comparable variance, normalized scores enable comparisons between the respective domain and the component scales. Higher scores represent better health-related quality of life.
The internal consistency of the SF-36 domains was evaluated using Cronbach's alpha. Values should exceed .70, but values ≥0.60 are acceptable in exploratory research .
Bivariate analyses used t-tests to compare means from the independent samples. Multivariable linear regression models were used to measure the impact of police officer characteristics on variation in the Physical Component and Mental Component Summary scores. All prediction variables were inserted as a block in each equation, using the default selection method ‘Enter’. Cases presenting studentized residual analysis varying around ± 3.000 standard deviations were identified as outliers. Data were analysed using SPSS version 20.0 (IBM Corp., Armonk, NY, USA).
All randomly selected policemen participated voluntarily in the survey. After explaining the objectives, participants signed a free informed consent form. All information related to the participants was confidential. The study was approved by the Ethics Committee in Research on Human Beings of the Medical School of Bahia for opinion number 554,724.