Indicators to assess the quality of programs to prevent occupational risk for tuberculosis: are they feasible?

Abstract Objective: to analyze the feasibility of quality indicators for evaluation of hospital programs for preventing occupational tuberculosis. Method: a descriptive cross-sectional study. We tested indicators for evaluating occupational tuberculosis prevention programs in six hospitals. The criterion to define feasibility was the time spent to calculate the indicators. Results: time spent to evaluate the indicators ranged from 2h 52min to 15h11min 24sec. The indicator for structure evaluation required less time; the longest time was spent on process indicators, including the observation of healthcare workers' practices in relation to the use of N95 masks. There was an hindrance to test one of the indicators for tuberculosis outcomes in five situations, due to the lack of use of tuberculin skin test in these facilities. The time requires to calculate indicators in regarding to the outcomes for occupational tuberculosis largely depends upon the level of organizational administrative structure for gathering data. Conclusions: indicators to evaluate the structure for occupational tuberculosis prevention are highly feasible. Nevertheless, the feasibility of indicators for process and outcome is limited due to relevant variations in administrative issues at healthcare facilities.


Introduction
Tuberculosis continues to be a threat worldwide.
Consequently, many healthcare workers (HCW) are at risk of being infected and acquiring this disease (1) . Good prevention programs should be established to avoid this undesirable outcome in healthcare facilities. To evaluate such programs, quality indicators can be used to identify the level of compliance for recommended practices.
Quality indicator technology has been increasingly used for evaluating health care practices. They are quantitative measures of features or attributes of a given process or system (2) , which may indicate the heath care quality provided, as well as specific needs for improvement (3) . Three classical categories have been used for their classification: structure, process and outcome (2,4) .The advantage of one over the other lies in the characteristics of the phenomenon to be measured.
Structure indicators refer to the features required, such as human resources, equipment, information systems, etc. Process indicators measure the dynamics of a given process, or how this particular process was performed. Outcome indicators measure the frequency in which event occurs, and assess final goals, such as mortality, morbidity or patient satisfaction (2,5) . Ideal indicators include features such as acceptability, objectivity, effectiveness, reliability, feasibility and availability, communication, interpretability, reproducibility, context, sensitivity to change, efficiency, and comparability (6) .

In 2006, a group of researchers in Brazil constructed
and validated a set of indicators designed to evaluate the quality of programs for healthcare-associated infection (HAI) prevention, including occupational tuberculosis. They can also be used to gauge the extent to which the control of HAI differs between different institutions (5,7) . Although the content was validated by professional experts, these indicators have not yet been fully tested.
Due to great difficulty in finding patterns for feasibility assessment in the literature, the best criteria for defining feasibility was previously discussed in a focus group with specialists (6,8) . The criterion "time" was chosen as a way of classifying these indicators as feasible. Providing the extent of time spent in measuring the indicator is as short as possible, the indicator is considered feasible. The shorter the time, the lower the human resources expense, and the more likely the indicator is to be widely use.
In the present study we aimed to analyze the feasibility of these quality indicators for the evaluation of programs for preventing occupational tuberculosis.

Methods
This descriptive, cross-sectional study tested the feasibility of quality indicators aimed at evaluating elements of structure, process and outcome of occupational tuberculosis prevention programs in different healthcare facilities. Time required for the calculationof the indicators was assessed as a measure of feasibility.
The quality indicators are described in Figure 1, The data analysis was descriptive.

Many quality indicators have been proposed
in the literature, however few have been evaluated regarding their feasibility for application, which creates a gap between theory and practice. Nevertheless, the recommendation for their use is quite frequent. To our knowledge, the present study is the first to evaluate the feasibility of quality indicators, using as the criterion the time spent on administering / calculatingthem.
Information on quality of care depends upon data availability. Therefore, quality is difficult to measure without correct and consistent information, which is often unavailable (8) . A previous study evaluated the feasibility of quality indicators related to radical prostatectomy and concluded that indicators not obtaining more than 25.9% of the necessary information were considered unenforceable (9) . It has also been previously shown that quality indicators for antibiotic treatment of complicated urinary tract infections were considered feasible if the data necessary to score the indicator can be abstracted from the available data for >70% of cases (10) . Indicators should require ease of obtaining data or ease of availability of the data as a condition of feasibility, resulting in minimal effort and additional cost (6,11) . It is a matter for discussion as to why, despite recommendations, some healthcare facilities in Brazil are not using the TST routinely, as we demonstrated in our sample. As an outcome to be measured, it was shown that the indicator for skin conversion (TOSCI) was not feasible due to this lack of compliance. The

Centers for Disease Control recommends the use of the
TST whenever there is the possibility of high exposure to tuberculosis (12) . HCWs should be periodically screened for latent tuberculosis infection using TST. As pointed out, concerning the healthcare facilities, many operational issues can interfere in the process. Among these issues, are the high turnover of HCWs, the limitations of the TST interpretation, and a potential booster effect of the BCG vaccine (13)(14)(15) .In order to overcome the booster effect, a two-step TST has been suggested in the literature (15)(16)(17) .
The TST has a high sensitivity, but lacks specificity in a vaccinated population, such as the HCWs in Brazil. Due to this feature, countries such as France and Japan are now recommending, with some restrictions, the gamainterferon release assays as a substitute for TST (18)(19) .
To note, in our sample, none of the healthcare facilities that were not using TST provided any other screening measure as a substitute.
The main outcome indicator (TOI), which measures the incidence of cases of tuberculosis among HCW, is quite simple to obtain, provided the Occupational Medicine Service has a structured form to record such cases. Usually cases of occupational tuberculosis are not as frequent as to warrant a great deal of effort in recording them. Besides this, the number of exposed HCWs is, in general, quite steady and does not require a sophisticated system to collect the information.
Despite this, many healthcare facilities are not aware of monitoring the annual incidence of occupational cases of tuberculosis.
The World Health Organization (WHO) shows that tuberculosis mortality in Brazil in 2013 was 3.2/100.000 and the prevalence was 57/100.000 (20) . Some authors have published similar results. A Peruvian study found a tuberculin test conversion incidence in medical students of approximately 3% (21) . A Brazilian study conducted in Belo Horizonte, MG, Brazil, where the tuberculosis incidence rate is 23/100.000, had the cooperation of 251 HCWs. The TST conversion was 5.1%, with the risk of infection of 1.4 (22) . A study aimed to identify the TST conversion rate of HCWs with previously negative TST results who had been working for less than 1 year in a hospital in Botswana, where tuberculosis is highly endemic. This population had a conversion rate of 4.2% for the entire group studied, or 6.87 per 1000 personweeks (23) .
A Chinese study showed that the health care workers' annual tuberculosis notification rates were lower than the general population. Healthcare workers with tuberculosis were a mean of 35.5 years old, with females out numbering males (58.0%>42.0%). The proportion of pulmonary tuberculosis was significantly higher among the women compared with men (88.5%>83.4%, p = 0.031). This study suggested that the priority for tuberculosis prevention in healthcare institutions should be given to the young female HCWs (1) .

An Argentinean study that included 15,276
HCWs from 15 centers found a mean incidence rate of tuberculosis in 111.3/100,000 HCWs (24) ; A Brazilian study demonstrated incidence rates in the general population of approximately 62/100,000, a prevalence of tuberculosis infection in HCW of 63.1% and an annual rate of tuberculin conversion of 10.7% (25) . In such an epidemiologic context, monitoring the incidence of occupational tuberculosis and the TST conversion can aid institutions in planning and evaluating strategies for occupational tuberculosis prevention, as demonstrated by other authors (13,15) .  (20) . For this control, it will be necessary to were not many objective criteria found in the literature that allowed for the evaluation of the applicability of indicators, so it was decided to use time as a marker.
However, we understand that this is a specific perspective that limits the study.
This study brings new insight to the applicability of previously validated quality indicators, revealing that even a validated indicator may not have all the properties of applicability; this approach needs to be considered to suggest recommendations for their use.
Moreover, strengths in the structure assessment, and weaknesses in the process and outcomes assessments, have been identified. Areas to be improved include maintaining periodic screening for latent tuberculosis using TST, monitoring the annual incidence of occupational cases of tuberculosis, and evaluating compliance with occupational prevention.

Conclusion
The indicators to evaluate the structure for occupational tuberculosis prevention are highly feasible.
The feasibility of applying indicators for process and outcome is limited, due to relevant differences in administrative issues at healthcare facilities, such as the system for data archiving and management.