The financial burden of an influenza outbreak in a small rehabilitation centre

Abstract We report an influenza outbreak in a 75‐bed rehabilitation centre and present the detailed microeconomic impact that it had during the season 2016/2017. The direct medical, direct non‐medical and indirect costs were calculated. The outbreak included 18 patients with influenza and 8 contact patients, leading to 86 days with isolation precautions. During the outbreak month, 25 (15%) employees were absent from work for 89 days (mean 3.6 days, SD ± 1.8), and during the entire influenza season 33 for 175 (5.3 ± SD 4.6) days, respectively. The economic burden related to the outbreak was 114 373 CHF (106 890 €, 112 131 $).


| INTRODUC TI ON
Seasonal influenza results in considerable morbidity and mortality, as well as increased healthcare utilization and costs. 1 Each year, the disease leads to substantial losses in productivity for both patients and caregivers. 2 The economic impact of influenza is of interest at both the microeconomic and the macroeconomic level. 3 Although several studies have calculated the cost-effectiveness of vaccination programs 4 or applied modelling tools, 5 precise data on the direct and indirect costs of influenza for a single institution are scarce. Here, we report an influenza outbreak in a rehabilitation centre and present its detailed microeconomic impacts.

| The outbreak
The institution Reha Chrischona (Bettingen, Switzerland) consists of 75 beds (one-bed or two-bed room [s]) and provides between 19 000 and 20 000 days of nursing care each year. 6 Specialized in the field of internal medicine, musculoskeletal rehabilitation and oncology rehabilitation, the institution employed 169 individuals in 2017. Standard precautions are applied to the care of all patients. Routine influenza vaccination is offered to healthcare workers without a mandatory vaccine policy. Influenza A subtype H3N2 was the predominantly circulating virus in the influenza season 2016/2017, the time interval defined to be from December 2016 to February 2017. 7 The outbreak in the institution occurred in January 2017.

| Outbreak management
Influenza outbreak management was implemented as described elsewhere, 8 after the detection of one confirmed influenza case and four patients with an influenza-like illness (ILI). An ILI case definition was implemented and included acute onset of new symptoms (ie, within 24 hours). The presence of fever (≥38°C) and cough, or fever and other respiratory illness symptoms or fever without a reasonable alternative diagnosis was required to fulfil the case definition. 9 A rapid influenza diagnostic test (polymerase chain reaction [PCR], Xpert Flu, Cepheid) was performed in nasopharyngeal swab specimens obtained from ILI cases. The influenza test result was communicated to both infection control specialists and healthcare personal of the institution within 24 hours. Daily active surveillance for acute respiratory illness among all patients and healthcare personnel was started and continued for 18 days after the last laboratory-confirmed influenza case was identified. Combined droplet-contact isolation precautions were applied to all patients who fulfilled ILI case definitions. The importance of standard precautions was stressed daily, and all healthcare personnel and visitors were requested to wear surgical masks and to perform hand hygiene before and after touching patients (irrespective of whether or not they had contact with an ILI case). Patients who fulfilled ILI case definitions remained in their private room or were discharged at home, depending on their health status and completion of rehabilitation programme. If patient transport to another institution was necessary, patients had to wear surgical masks during the transport and information on the influenza test result was communicated.
Isolation precautions in confirmed influenza cases were applied for at least five days after illness onset (at least ten days in immunocompromised patients) or until 24 hours after the resolution of fever and respiratory symptoms, whichever was longer. Oseltamivir (75 mg) was administered twice daily to all patients who fulfilled the ILI case definition and was continued in laboratory-confirmed cases for five days or in immunocompromised patients for ten days.
PCR was repeated in immunocompromised patients with confirmed influenza after five or ten days of antiviral treatment. Patients exposed to a PCR-positive influenza patient in the same patient room were pre-emptively treated with oral oseltamivir (75 mg twice daily) for five days. Antiviral chemoprophylaxis was only offered to the defined patient group but not to all individuals.

| The economic burden
Data (direct costs and number of absentees from work per day during the outbreak) for the calculation of the economic burden were collected prospectively for this study. To calculate the economic burden, we estimated the direct medical costs, direct non-medical costs and indirect costs. 3 We defined direct medical costs as those associated with diagnostics and treatment of patients with ILI and their room neighbours (ie, contact patients). Direct non-medical costs included the extra expense for droplet-contact isolation precautions. These costs were calculated in a previous investigation and extrapolated for this study. 10 They included costs for extra materials used (eg, surgical masks, gloves, gowns), increased workload and off isolation activities specifically for patients under isolation precautions. 10

| RE SULTS
The number of patients with contact-droplet isolation precautions per calendar day during the outbreak month is summarized in Figure 1A. Eighteen patients had laboratory-confirmed influenza, and eight contact patients were identified. Two patients with laboratory-confirmed influenza were referred to a tertiary care centre because of decompensated heart failure, after one day and five days with isolation precautions, respectively. Five patients with laboratory-confirmed influenza and one patient with ILI and negative influenza test result were discharged at home, because they were in good health and their rehabilitation programme was completed.
Their periods with isolation precautions ranged from one day to five days prior to discharge. In total, 86 isolation precautions days were implemented in 18 days. An index case for the nosocomial outbreak was not identified. None of the patients died. No new ILI case was observed after 7 days. The numbers of absentees per day during the outbreak month are shown in Figure 1B. In January 2017, TA B L E 1 Self-reported work loss because of ILI during influenza season 2016/2017 at a rehabilitation centre F I G U R E 1 A, The number of patients with contact-droplet isolation precautions per calendar day during the outbreak. B, Number of absentees per day during the outbreak month. The caregiver absent at January 2nd was not included in the study, because absence was more than 7 days prior to the start of the outbreak the employer reported that 25 (15%) employees were absent for 89 days (mean 3.6 days, SD ± 1.8, range 1-8 days per employee).
The day with the highest numbers of patients with isolation precautions (16 January 2017; 12 individuals, Figure 1A) coincided with the day of the highest number of absentees from work (11 individuals, Figure 1B).
Of the distributed questionnaires (169), 158 were returned and 150 (89%) were completed and included in the study. Hundred twenty (80%) of 150 employees were caregivers with patient contact, and 90 (60%) were exposed to patients with influenza.

| D ISCUSS I ON
The rapid infection control intervention was effective for the de-  in previous studies. 12 Differences in the circulating influenza strains and the proportion of vaccinated individuals may influence the range of absence days because of ILI. Vaccine effectiveness against seasonal H3N2 in working-age adults was reported to be 35%. 13 We are unable to evaluate whether there was a shorter duration of work absence for vaccinated patients who contracted influenza than for unvaccinated employees. Nonetheless, the proportion of vaccinated employees was low (15%). Policy change and constant education for healthcare personal may potentially increase this proportion.
The study has limitations. We did not encounter the numbers of surgical masks and the volume disinfectant for hand hygiene used by visitors and personnel without patient contact. Thus, the direct nonmedical costs are likely underestimated. The productivity loss data should be interpreted with caution, because there may be a recall bias of self-reported absence and the anonymized questionnaire was not previously validated. Also, these data need to be analysed from a seasonal perspective over several years. Nonetheless, in this investigation, we demonstrated the financial damage at a single institution caused by an outbreak. In light of the cost pressure to which rehabilitation centres are exposed, these data may have an implicit effect on influenza prevention strategies. These include programmes to increase the vaccination rates among healthcare personnel and protocols for antiviral chemoprophylaxis. A detailed and rapid outbreak management concept against influenza may turn to be cost-effective, in particular for small institutions with limited resources.

ACK N OWLED G EM ENTS
We thank René Gröflin (CEO), Gabriela Pflauma (RN, head nurse) and all employees of Bürgerspital Basel, Reha Chrischona, Bettingen, Switzerland, for their help during the outbreak and for providing data for the generation of this manuscript.