Attitudes towards preventive tuberculosis treatment among hospital staff

Study setting and population

The study was performed at Liverpool Hospital, a tertiary hospital with a major TB clinic in Sydney. Liverpool Hospital TB clinic treats approximately 10% of all TB patients in Australia and 20% of all patients in the state of New South Wales (NSW). In this setting, the TB incidence is 7.3 per 100,000 population per year (Barry et al., 2012).

Since the introduction of a new policy directive on HCW screening in NSW in 2011 new HCWs must only undergo TB screening if they were born in a country with a high incidence of TB, or if they have travelled to a country with a high incidence of TB for three months or longer (NSW Health, 2011). The state policy directive defines a country with a high incidence of TB as one that has a TB incidence ≥60 cases per 100,000 population per year. Existing staff members who lived in a country with a high incidence of TB for more than 3 months within the last 3 years must also undergo TB screening.

During the screening process, a medical history is taken; a tuberculin skin test (TST) is performed and a chest X-ray is done if the TST is positive. The results of these investigations are then reviewed by a doctor at the TB clinic, who decides whether the HCW is booked for a doctor’s appointment or not. When there is no evidence of active TB, the doctor will decide whether to offer preventive treatment to a HCW with evidence of LTBI (usually TST ≥10 mm) based on an assessment of the individual’s risk of developing TB. If there is concern about a possible false positive TST result (e.g., in a HCW who received a Bacillus Calmette-Guérin (BCG) vaccination, which is known to cause a TST cross-reaction), the doctor may also decide to order an additional blood test (Interferon Gamma Release Assay (IGRA)) to confirm true LTBI.

Study design

A paper-based questionnaire was sent to randomly selected nurses and doctors, attached to the fortnightly pay slips. Respondents were asked to return the completed questionnaire anonymously in the attached return envelope. Randomisation was performed using a random number generator in Microsoft Excel. Additionally, one of the researchers (VP) approached doctors and nurses in person using a random starting point from a list of hospital locations (including wards, outpatient clinics and clinical meetings) and then approaching every second HCW. Staff who were approached in person were asked if they had received a questionnaire with the pay slip. If they indicated that they had not yet received a questionnaire, they were given a questionnaire and asked to return it anonymously in the attached envelope. They were asked to fill in only one questionnaire should they later discover a mailed questionnaire at home. We did not have a formal control mechanism to exclude duplicate submissions from the same person. The risk for duplicate submissions was, however, perceived to be very low because there was no financial incentive attached to providing a response.

Questionnaire

The questionnaire for hospital staff (see supporting information) included questions on HCWs’ demographic background including their age, gender, profession, hospital department, level of patient contact, country of birth, and history of overseas stay. Additional questions addressed the personal experience with TB screening including history of BCG vaccination, results of screening chest X-rays, results of TST and IGRA tests and knowledge about their personal TB status (no evidence of LTBI, evidence of LTBI, active TB or past medical history of TB). Participants were asked whether they were offered treatment for LTBI and, whether they accepted treatment for LTBI. They were questioned about their general opinion on treatment for LTBI.

Definitions

A TB incidence of ≥60 per 100,000 population was chosen to define a country with a high incidence of TB, as per the NSW Health Policy Directive (NSW Health, 2011). Incidence rates of TB were based on estimates of the World Health Organization (WHO) accessed on 01/10/2013 (World Health Organization, 2013). A positive screening test result was defined as having a positive TST (≥10 mm) and/or IGRA. Staff working in respiratory medicine was defined as HCWs whose work focused on the care of patients with respiratory problems, not just staff involved in the care of TB patients. The speciality of respiratory medicine was highlighted in the study because the TB clinic at the study hospital was integrated into the department of respiratory medicine.

Statistical analyses

Calculation of proportion of responses was based on the total number of HCWs who answered a specific question in the questionnaire, in other words, the denominator for each question was the number of staff who answered the question by ticking one of the boxes. Data were extracted from the paper-based questionnaire and entered into a Microsoft Excel database.

Logistic regression and odds ratios (ORs) with 95% confidence intervals (95% CIs) were used to assess associations between demographic characteristics and survey responses. The independent effect of potential predictors of survey responses among HCWs was estimated using multivariable logistic regression for binary outcomes, in which the dependent variable could take only two values, yes/no. Responses that indicated that the respondent had no opinion or did not know the answer to the question were thus excluded for this analysis. Multivariable analysis included adjustment for age, sex, profession, specialty, country of birth, screening results, being offered treatment and history of BCG vaccine. Statistically significant results were defined as p < 0.05 SPSS Statistics v21 (IBM Corp, Armonk, NY, USA) was used for the statistical analysis.

Ethical approval & consent

The study protocol was approved by the South Western Sydney Local Health District Ethics Committee (project number 12/265). Attached to all questionnaires was a participant information sheet which included the following information regarding participant consent: “Your participation in this survey is entirely voluntary. You may choose not to take the survey, or to skip any questions that you do not want to answer. Your completion of the questionnaire serves as your voluntary agreement to participate in this research project.”

Participants

A total of 733 questionnaires were sent to randomly selected nurses and 431 to randomly selected doctors; an additional 87 and 53 questionnaires were directly handed to nurses and doctors respectively. Of 1,304 HCWs who received a questionnaire, 318 responded, resulting in a response rate of 24%. Seven responses were excluded because the questionnaires were blank, so 311 responses were included in the analysis. The response rate was 22% (105/484) among doctors and 25% (206/820) among nurses. Forty-seven percent (144/307) of respondents were Australian-born. Ninety-five percent of respondents (295/311) had undergone TB screening, including 151who had undergone screening at the hospital’s TB clinic. The majority (225) had undergone screening prior to February 2011, when a new screening policy for HCWs was introduced in NSW, limiting TB screening to HCWs considered to be at risk of previous TB infection. Table 1 summarizes the demographic characteristics of all HCWs who participated in the study.

Experience with screening for latent TB infection at the hospitals’ TB clinic

Of the 295 HCWs who underwent screening, 258 (87%) reported having undergone TST and/or IGRA testing (Fig. 1). Of those who underwent a TST and/or IGRA test, 116 (45%) indicated that they had a positive screening test result suggestive of LTBI, of which 10 (9%) did not answer the question on whether they were offered treatment for LTBI or not. Only 13% (14/106) of HCWs with evidence of LTBI indicated that they were offered preventive TB treatment; 40% (42/106) were not offered treatment, and 47% (50/106) felt that this question did not apply to them. Sixty-four percent of HCWs (9/14) accepted the offered preventive TB treatment. HCWs who refused treatment felt that they could monitor signs and symptoms of TB themselves (3/5); stated they feared side effects (2/5), felt that the treatment was unnecessary (2/5) and stated that they did not need treatment as they had been vaccinated (1/5).

Attitudes towards preventive TB treatment

Of the respondents who expressed an opinion on the topic, 74% (164/223) supported the idea that hospital staff who had evidence of LTBI should receive preventive treatment while 26% (59/223) did not think that preventive TB treatment was indicated in this situation. A further 80 respondents did not express any opinion on the topic. In bivariate logistic regression analysis there was a positive association between support for preventive TB treatment for employees in non-respiratory specialties (OR 7.83, 95% CI [2.27–22.51] for staff in medical specialties other than respiratory medicine and OR 20.83 95% CI [4.53–95.89] for surgical staff, compared to respiratory staff). There were also associations between support for preventive TB treatment and female gender (OR 1.94, 95% CI [1.05–3.61]), nursing staff (OR 2.36, 95% CI [1.28–4.34]) and positive TST status (OR 2.41 95% CI [1.21–4.82]). In multivariable logistic regression, only the respondent’s specialty had a significant association with a favourable attitude towards preventive TB treatment. Staff from non-respiratory specialties were more supportive of preventive TB treatment than respiratory staff (aOR 8.62, 95% CI [2.69–27.61] for staff in other medical specialties and aOR 16.94 95% CI [3.22–89.05] for surgical staff) (Table 2).

Eighty-one percent (198/244) of respondents were in favour of being offered preventive TB treatment if they had evidence of LTBI personally and 19% (46/244) preferred not be offered preventive TB treatment personally. Another 53 respondents had no opinion on the topic. There was greater support for personal preventive TB treatment among staff from non-respiratory staff compared to respiratory staff (85%, 187/221 versus 48%, 11/23 respectively (Table 2)). The specialty hospital staff worked in was the only characteristic that was significantly associated with the attitude towards personal preventive TB treatment in bivariate and multivariable logistic analyses.

The majority of respondents (70%, 185/264) stated they would require more information to decide whether they personally would want to receive preventive TB treatment if they had evidence of LTBI, 30% (79/264) felt that they were sufficiently informed and a further 31 had no opinion on the topic.

The majority of respondents who expressed an opinion (73%, 200/273) thought that Australian born HCWs without significant overseas stay should be screened for LTBI; 26% (73/273) thought that they should not be screened. A further 34 respondents voiced no opinion. A third of respiratory staff (9/27) believed that these HCWs should be screened, compared to 78% (191/246) of those in other specialities.

Education and knowledge

Twenty-one percent of respondents (64/306) indicated that they did not know the difference between active and latent TB. Only 69% of respondents aged 30 years or less thought that they knew the difference compared to 89% of those in the 31–40 years age group (aOR 3.32, 95% CI [1.23–8.99]) (Table 3). Nurses were less likely to indicate that they knew the difference than doctors (69%, 138/201 versus 99%, 104/105; aOR 0.033, 95% CI [0.004–0.26]). Univariate analysis revealed that women were significantly more likely than men to indicate that they did not know the difference between the two conditions, but this was not confirmed in multivariable analysis (Table 3).

Of the HCWs who had undergone TST and/or IGRA testing only 33% (75/230) felt adequately informed about the meaning of their test results, 16% (37/230) did not feel adequately informed, and 51% (118/230) were not even aware that the question on information about test results applied to them.