Case-Patient Description

On July 2, 2008, after returning from a visit to Uganda from June 5 through June 28, a 41-year-old woman showed development of chills and high fever. She was admitted to hospital A on July 5. Initially, hemorrhagic fever was not included in the differential diagnosis and she was placed in a general ward among other patients, without specific contact precautions. Because she later showed clinical deterioration, liver failure, and tendency to hemorrhage, hemorrhagic fever was suspected and she was transferred to hospital B on July 7. In hospital B, she was placed in strict isolation in accordance with guidelines prescribed for pathogens belonging to Hazard Group 4 (18). We placed the patient in a single room with negative air pressure ventilation and an anteroom for 2 reasons. First, although evidence for airborne transmission of MHF in humans has not been documented, transmission by aerosols has been demonstrated in animal models (7). Second, the patient was likely to undergo aerosol-generating procedures (e.g., endotracheal intubation) while in the latter stages of illness when viral loads in body fluids were expected to be high.

On July 10, the diagnosis of MHF was confirmed; the next day, the patient died. A complete case history and the public health response have been reported (1).

Control Measures

The outbreak response team formulated measures for follow-up of contacts considered to be at risk for exposure. Measures were based on preexisting national and international guidelines on management of hemorrhagic fever caused by filoviruses (18–20). The patient was considered to be potentially infectious from the onset of fever (July 2) until death (July 11). The period of monitoring contacts was set at 21 days after the last contact with the patient or patient body fluids. The public health service traced the contacts in the community. The hospital hygienist and occupational physician, and attending physicians were responsible for in-hospital contacts. Contacts were provided with written instructions (Table 1).

Contacts classified as at high risk for contracting the disease (i.e., had unprotected contact with the patient or her body fluids) were asked to measure their temperature twice a day and report it to the health care provider. Furthermore, they were prohibited from leaving the country and were told to report any intention to leave to the public health authority. Contacts were asked to notify the assigned health care provider immediately if they had fever (body temperature >38°C measured twice at least 12 hours apart) or any abnormal symptoms (e.g., vomiting, headache, abdominal pain, diarrhea, jaundice). Contacts who had handled the patient or her body fluids while carrying out strict isolation measures were perceived to be at low risk for exposure. They were asked to measure their temperature twice a day, but only to report a temperature >38°C. This group was strongly encouraged not to travel abroad.

Study Population

A retrospective cohort study involving the 130 contacts was conducted by using an online questionnaire. High-risk contacts (n = 64) included household contacts, health care providers at the general practice, those involved in patient care at hospital A (nurses, physicians, and laboratory workers), patients sharing the same hospital ward in hospital A, and cleaning staff. Low-risk contacts (n = 66) included health care and laboratory workers at hospital B who had all taken appropriate personal protective measures. Data were collected from December 2008 through February 2009, which was 5–7 months after possible exposure. At the same time, serologic testing was conducted to assess asymptomatic transmission.

Variables and Instruments

Statistical Analysis

Data were analyzed by using SPSS version 18 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were calculated. Means and SDs were calculated for answers given on the Likert scale. Results were stratified by type of exposure risk (i.e., high or low risk for disease transmission). Differences in proportions were calculated by using the χ² test. Differences in means were calculating by using the Student t test. A p value <0.05 was considered significant.

We constructed overall scales. The compliance with measures scale (Cronbach α 0.71) included 4 questions regarding temperature monitoring, temperature reporting, and prohibition on travel. The interference scale (Cronbach α 0.64) included 11 questions on perceived restrictions on social life, difficulties with control measures, extra costs, and anxiety of contacts and their families. The clarity of instructions scale (Cronbach α 0.82) included 5 questions on explicitness, completeness, unambiguity, confusion, and redundancy of the provided instructions. The overall scale for understanding of measures (items regarding awareness of the measures and their rationale) proved invalid and was not used in further analysis.

To determine which variables influenced stress levels during the monitoring period, we developed a linear regression model by using the IES during monitoring as a dependent variable and personal characteristics (sex, education, age), risk level, clarity of instructions scale, compliance scale, and interference scale as independent variables. The same model was constructed by using the IES after the monitoring period as the dependent variable.

Serologic Testing

Serologic testing was performed by using an immunofluorescence antibody (IFA) assay with blood samples collected from contacts 5–7 months after the monitoring period. IFA test slides were prepared at the Bernhard Nocht Institute for Tropical Medicine (Hamburg, Germany) by using the patient strain. Laboratory methods have been reported (1).

Response

Of 130 eligible participants, 78 (60.0%) completed the questionnaire. One person provided systematically inconsistent answers and was excluded, which left 77 respondents (59.2%) for statistical analysis. The response rate was 70.3% (45/64) in the high-risk group and 48.5% (32/66) in the low-risk group (Table 2). Of the 77 respondents, 46 (59.7%) were female, 44 (57.1%) did not have children in the household, and 11 (14.3%) lived alone. Mean age was 38 years in the high-risk group and 43 years in the low-risk group (this difference was not significant). The high-risk group had more women (33/45, 73.3%) than the low-risk group (13/32, 40.6%) (p = 0.004) and had a lower education level (Table 2). Respondents were comparable to nonrespondents with respect to sex (M:F ratio 40.0%:60.0% vs. 42.0%:58.0%). During the monitoring period, nonspecific symptoms (those commonly occurring during the prodromal phase of hemorrhagic fever syndromes, e.g., headache, malaise, fatigue) (2) developed occasionally in some contacts, but no significant differences were observed between the risk groups.

Understanding Measures

Of the 77 respondents, 60 (77.9%) were aware of the request not to leave the country, and 17 (22.1%) chose the neutral option. Of 60 contacts aware of the request, 54 (90.0%) in the high-risk group agreed with this request (p = 0.04). Reasons given for negative answers were no direct contact with the patient and thus no risk (n = 2), imperative reasons to leave the country such as illness in the family (n = 2), travel restrictions were meant for others, not for me (n = 1), and borders do not stop diseases (n = 1). All respondents were aware of the need to measure temperature twice a day, 67 (87.0%) agreed with the necessity of measuring and reporting temperature, and 7 (9.1%) had no opinion. Not feeling ill was the reason for disagreement in the remaining 3 (3.8%) respondents. Only 58 (75.3%) of the 77 respondents correctly identified the rationale of temperature monitoring. Respondents in the high-risk group agreed more often than their low-risk counterparts with the necessity of daily measuring and reporting for temperature (mean ± SD of agreement in the high-risk group 4.7 ± 1.0 vs. 1.6 ± 1.3 in the low-risk group) (p<0.001). Written instructions with detailed information on the control measures were received by 61 (79.2%) of 77 respondents. Of these 61 respondents, 45 (73.7%) found this information to be clear, 40 (65.6%) complete, 37 (60.6%) unequivocal, 4 (6.5%) confusing, and 3 (5.0%) redundant. There were no significant differences between the risk groups.

Compliance with Measures

Of 45 respondents who were prohibited from traveling (high-risk group), 17 (37.7%) had planned a holiday trip abroad, of whom 12 (70.6%) cancelled the holiday trip during the surveillance period, and 1 (5.9%) postponed the trip. Two (11.8%) high-risk contacts were already abroad when the diagnosis in the index case was made and 2 (11.8%) other contacts left the country a few days before the end of the surveillance period, despite prohibition on travel. Risk management in these persons, including communication with government authorities in countries to which they traveled, has been reported (1). Temperature was monitored twice a day by 62 (80.5%) of 77 persons, and 75 (97.4%) of 77 had a thermometer at their immediate disposal. Strict compliance with daily temperature monitoring decreased from 80.5% (62/77) in week 1 to 66.2% (51/77) in week 3. Differences per risk group per week in reported compliance with temperature monitoring and reporting are shown in Table 3.

Interference of Measures with Daily Life

The prohibition on leaving the country was perceived as difficult by 18 (23.4%) of 77 respondents; 11 (14.3%) perceived serious restrictions on their social life, and 24 (31.2%) reported feeling stressed on a regular basis during monitoring. Daily temperature monitoring was believed to be troublesome by 25 (32.5%) of 77. Extra costs were involved for 19 (24.6%) of the contacts. The questionnaire did not ask for specification of costs incurred. Of 47 health care workers among respondents, 13 (27.6%) intensified adherence to infection prevention guidelines at work during the monitoring period (mean ± SD intensified adherence in the high-risk group 1.3 ± 0.4 vs. 1.03 ± 0.2 in the low-risk group) (p = 0.006). Sustained and intensified adherence to infection prevention at work after the monitoring period ended was reported by 10 (21.3%) of 47.

Being identified as a contact caused anxiety in respondents, as reflected in the high percentage who were afraid of contracting MHF (45/77, 58.4%), those who were concerned about infecting other members of their household (31/77, 40.3%), or those who were afraid that a colleague might have been unknowingly infected by the index patient (35/77, 45.5%). Of 77 respondents, 28 (36.4%) reported that their family members were disturbed by control measures. Furthermore, 31 (40.3%) of 77 reported that their family members expressed anxiety about becoming infected, and 25 (32.5%) of 77 reported that their partners were disturbed by restrictive control measures. The results on the overall interference scale were compared and showed that control measures caused more interference with daily life (lower mean ± SD) in the high-risk group (mean ± SD 1.5 ± 0.3 vs. 1.8 ± 0.2; p<0.001) than in the low-risk group.

Impact of Event Scale

Figure

Figure. . . Distribution of individual scores on the impact of event scale (IES) during and after (a 7-day period before completion of a questionnaire) the monitoring period among contacts of the...

Means were calculated on the basis of the answers on the IES and subscale domains (intrusion, avoidance, and hyperarousal). Overall psychological distress was measured by the mean ± SD score on the IES during monitoring (2.8 ± 2.6, range 0.2–10) and 5–7 months after monitoring 0.9 (0.9 ± 1.7, range 0–9.5; p<0.001). Despite the low overall mean score, high individual scores were observed (Figure). On the subscale domains, the mean scores were higher during the monitoring period than after it, and the highest scores were reported in the high-risk group.

In the linear regression model, the score on the interference scales was a predictor of the score of the IES during monitoring, which measured the psychological effect of this event (p<0.001). After monitoring ended, the remaining psychological effect measured by the IES score was influenced by the degree of interference (p<0.001) and level of education (p = 0.015); persons with a lower education level reported more psychological consequences (Table 4).

Serologic Testing

IFA serologic testing was conducted in 85 (65.4%) of 130 participants, which represented 50 (78.1%) of 64 in the high-risk group and 35 (53.0%) of 66 in the low-risk group. All serum samples were negative for IgG and IgM against Marburg virus.