JAMDA 15 (2014) 226.e1e226.e6
JAMDA journal homepage: www.jamda.com
Review
Effectiveness of Influenza Vaccination in Institutionalized Older Adults: A Systematic Review Tuen-Ching Chan MBBS, FHKAM a, b, *, Ivan Fan-Ngai Hung MD, FHKAM c, James Ka-Hay Luk MBBS, FHKAM a, Leung-Wing Chu MD, FHKAM b, Felix Hon-Wai Chan MB Bch, FHKAM a a b c
Department of Medicine and Geriatrics, Fung Yiu King Hospital, Hong Kong SAR, China Division of Geriatrics, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong SAR, China Division of Infectious Disease, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong SAR, China
a b s t r a c t Keywords: Influenza vaccination efficacy institutionalized older adults review meta-analysis
Introduction: Influenza infection is common among institutionalized older adults. Many nonrandomized observational studies on influenza vaccination suggested that it could reduce influenza-related hospitalizations and mortality in institutionalized older adults. Criticism regarding the effectiveness of influenza vaccine estimated by nonrandomized observational studies include the frailty selection bias and use of nonspecific outcome, such as all-cause mortality. Methods: We conducted a systematic review of studies of influenza vaccination in institutionalized older adults to determine the effects on clinical outcomes. We searched for studies from 3 databases from 1946 to June 2013 assessing effectiveness against influenza infection. We selected studies with good comparability between vaccine group and control group. We expressed vaccine effectiveness (VE) as a proportion, using the formula VE ¼ 1erelative risk or 1eodds ratio. We focused on the following outcomes: influenzalike illness (ILI), laboratory confirmed influenza, hospitalizations due to ILI, or pneumonia and death due to influenza or pneumonia. We did not include all-cause mortality. Results: Eleven studies that satisfied the inclusion criteria were identified, representing 11,262 institutionalized older adults. After meta-analysis, we found a significant reduction in pneumonia (VE: 37%, 95% confidence interval [CI]: 18%e53%, P ¼ .001) and death due to pneumonia or influenza (VE: 34%, CI: 10% e53%, P ¼ .01). There was no significant heterogeneity between studies. There was no significant publication bias. Conclusion: Influenza vaccination in institutionalized older adults could reduce pneumonia and death due to pneumonia or influenza. Influenza vaccination is recommended for institutionalized older adults. Ó 2014 - American Medical Directors Association, Inc. All rights reserved.
Influenza is a leading cause of infection death among older adults.1 The main strategy for prevention and control of seasonal and pandemic influenza for the past 50 years has been vaccination. The yearly influenza vaccination of at-risk individuals is a common practice, and adults 65 years and older (60 years in some countries) are suggested to have the vaccination.2,3 However, despite the widespread availability of influenza vaccination programs, vaccine coverage rates are generally unsatisfactory.4,5 This largely explains why influenza infections are still a major public health concern across the world.
The authors declare no conflicts of interest. * Address correspondence to Tuen-Ching Chan, MBBS, FHKAM, The University of Hong Kong, Division of Geriatrics, Department of Medicine, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong SAR, China. E-mail address: [email protected] (T.-C. Chan).
Institutionalized older adults have high infection-related hospitalization and mortality rates.6e9 High vaccine coverage in this population is very important. Most studies on influenza vaccination among older adults were nonrandomized observational studies. Previous systematic reviews on influenza vaccination suggested that influenza vaccination could reduce influenza-related hospitalizations and mortality in institutionalized older adults.10,11 However, the effectiveness of the influenza vaccine estimated by nonrandomized observational studies is frequently criticized for 2 major reasons: the frailty selection bias and use of nonspecific outcomes.12e14 Some physicians may feel that vaccine is futile in very frail older adults. A subset of undervaccinated frail elderly people could contribute a substantial proportion of adverse outcomes. If undervaccination were a direct consequence of the poor health status of those who are elderly, it would be the major source of bias. Many studies used nonspecific outcomes, especially all-cause mortality, that may contribute
1525-8610/$ - see front matter Ó 2014 - American Medical Directors Association, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jamda.2013.10.008
T.-C. Chan et al. / JAMDA 15 (2014) 226.e1e226.e6
to overestimation of vaccine efficacy.14 Therefore, we performed a systematic review and meta-analysis for effectiveness of influenza vaccination in institutionalized older adults with measures to minimize these confounding factors. Methodology We carried out a literature search using the following electronic databases: Medline 1946 to June 2013, Embase 1974 to June 2013, and the Cochrane Controlled Trials Register (The Cochrane Library 2013, Issue 4). We used the medical subject headings “influenza vaccine,” “influenza immunization,” “nursing home,” “institution,” “long-term care,” “elderly,” and “older adults.” Citations of relevant articles were reviewed. The Science Citation Index was used to trace any articles quoting a previous meta-analysis. Trial identification was based on PRISMA statement 2009.15 Studies with the following criteria were included: 1. Participants who were institutionalized adults aged 60 years or older. 2. Study design was a randomized controlled trial or observational study. 3. The outcome measures were all clinically relevant: clinically defined influenza-like illness (ILI), laboratory-confirmed influenza infection, pneumonia, hospitalization due to pneumonia or ILI, or death due to influenza or pneumonia. ILI was defined as fever together with respiratory and systemic signs and symptoms. Laboratory-confirmed influenza infection was defined as viral isolation with or without serological supporting evidence. Pneumonia was defined by clinical and radiological changes suggestive of pneumonia. We did not include all-cause mortality. 4. Data on outcomes of interest had to be reported in numerical format, including numerator and denominator data. Two observers independently extracted data and methodological details from the unmasked published reports. The following data were extracted: (1) description of the interventions, including the strain of the vaccine; (2) total number of subjects, and their age and sex; (3) follow-up duration; (4) study outcome definition and the number of outcome events corresponding to the requested definition in each group; (5) whether there was an outbreak in the nursing home during the study; and (6) whether the vaccine strain matched the circulating influenza strain. Discrepancies were documented and resolved by the consensus of both investigators. In a randomized controlled trial, Jadad’s16 system was used by the 2 reviewers to score each report. In Jadad’s16 system, 1 point is allocated for each of randomization, blinding, and description of withdrawals, and an additional point added for each of the well-described methods of randomization and blinding. Points were subtracted for clearly inadequate randomization or blinding. The maximum possible score was 5 points, and we took studies scoring less than 3 to be of low methodological quality. In nonrandomized observational studies, the Newcastle-Ottawa quality assessment scale was used by the 2 reviewers to score each study.17 In the Newcastle-Ottawa quality assessment scale, any study is graded according to 3 categories: selection, comparability, and outcome. There are 4 items in selection (representativeness of the exposed cohort/case control study; selection of the nonexposed cohort/case control study; ascertainment of exposures; and demonstration that outcome of interest was not present at start of study), 1 item in comparability (comparability of cohorts on the basis of the design or analysis), and 3 items in outcome (assessment of outcome; was followed long enough for outcomes to occur; and adequacy of follow-up of study). A study can
226.e2
be awarded a maximum of 1 star for each numbered item within the selection and outcome categories. A maximum of 2 stars can be given for comparability. In this review, 1 star would be given for comparability if age were compared between groups. Another star would be given if comorbidity and/or functional status were compared between groups in the study. The maximum possible score of Newcastle-Ottawa quality assessment scale was 9 points. We took studies scoring less than 8 to be of low methodological quality and only studies scoring 2 in the comparability category were included in the review. Election discrepancies were documented and resolved by the consensus of both observers. For each included study, if there were 2 or more nursing homes involved and outcomes of the 2 nursing homes were individually presented, data of each nursing home were individually analyzed during the meta-analysis (eg, Saito 2002a and Saito 2002b). Because in Saito 2002, there were 2 nursing homes in this cohort study. And the data of each individual nursing home was presented, hence the data of each nursing home was analyzed individually. Calculations were done with comprehensive meta-analysis software 2.0 (Biostat, Englewood, CA). The Mantel-Haenszel fixed effects model odds ratio (OR) was used to quantify the protective effect of the influenza vaccine. The heterogeneity between studies was assessed using the Cochrane Q statistical method, and a P value less than .10 was considered significant. When significant heterogeneity was found, a method based on a random effects model was used. Analyses were performed for different clinical outcomes. Within each clinical outcome, subgroup analyses were performed according to outbreak status in the nursing home and matching between vaccine strains and circulating influenza strain. There were 4 possible subgroup analyses for each clinical outcome: A, There was an outbreak, vaccine strain matched circulating influenza strain; B, There was an outbreak, vaccine strain unmatched with circulating influenza strain or unknown; C, There was no outbreak, vaccine strain matched circulating influenza strain; and D, There was no outbreak, vaccine strain unmatched with circulating influenza strain or unknown. Publication bias was assessed using funnel plot. We expressed vaccine effectiveness (VE) as a proportion, using the formula VE ¼ 1erelative risk or 1eOR.
Fig. 1. Flow diagram of study identification, based on the PRISMA statement.
226.e3
Table 1 Characteristics and Outcomes of Included Studies Study [Reference No.] Design
Sample Size
Outcome Measurement
Vaccine Matching
Quality Assessment Scales* Selection Comparability Outcome
Prospective cohort (1 y)
Currier 1988 [19]
Retrospective case-control for outbreak (6 mo)
Gross 1988 [20]
Prospective cohort study of outbreak (6 mo)
Horman 1986 [21]
Retrospective case control for outbreak (3 mo)
Leung 2007 [22]
Monto 2001 [23]
-Death due to pneumonia
Matched
****
**
***
-Clinically defined ILI -Clinically defined pneumonia
Not matched
***
**
***
-Laboratory-confirmed influenza -Clinically defined pneumonia
Matched
***
**
***
-Clinically defined ILI -Clinically defined pneumonia -Death due to ILI
Matched
****
**
***
3177 residents (2943 vaccinated and 234 controls) from 46 nursing homes
-Clinically defined ILI -Clinically defined pneumonia -Hospitalization due to ILI
Unknown
****
**
***
2351 residents (1728 vaccinated and 623 controls) from 26 nursing homes 128 residents (60 vaccinated and 68 controls) from a nursing home 1352 residents (678 vaccinated and 674 controls) from a nursing home 1629 residents (1074 vaccinated and 555 controls) from 20 nursing homes 84 residents (65 vaccinated and 19 controls) from a nursing home 97 residents (45 vaccinated and 52 controls) from a nursing home
-Clinically defined ILI -Clinically defined pneumonia -Death due to ILI -Clinically defined ILI -Hospitalization due to ILI -Death due to ILI -Clinically defined pneumonia -Death due to pneumonia
Matched
****
**
***
Matched
****
**
***
Comorbidity, including mainly cardiac and pulmonary disease
Matched in 1 y, **** other unmatched
**
***
-Clinically defined ILI
Matched
****
**
***
Comorbidity, including cardiac, pulmonary, stroke, DM, and so forth Comorbidity, including cardiac, pulmonary disease and stroke
-Clinically defined ILI -Death due to ILI
Matched
****
**
***
Comorbidity, including mainly cardiac and pulmonary disease
-Clinically defined ILI -Laboratory-confirmed influenza -Clinically defined pneumonia -Hospitalization due to ILI -Death due to ILI
Matched
****
**
***
Functional status using level of mobility
1859 residents (1214 vaccinated and 645 controls) from 58 nursing homes 121 residents (87 vaccinated and 34 controls) from a nursing home 305 residents (181 vaccinated and 124 controls) from a nursing home 159 residents (100 vaccinated and 59 controls) from a nursing home
Retrospective case control (1 y)
Prospective cohort study (6 mo)
Murayama 1999 [24] Retrospective case control for outbreak (6 mo) Saah 1986 [25]
Prospective cohort study (6 mo)
Saito 2002 [26]
Prospective cohort study (1 y)
Strassburg 1986 [27]
Retrospective case control study for outbreak (6 mo)
Taylor 1992 [28]
Retrospective case control study for outbreak (6 mo)
Comorbidity using comorbidity index; functional status using Barthel Index Comorbidity including cardiac and pulmonary disease; cognitive status Comorbidity, including cardiac, pulmonary, stroke, and so forth; medication Comorbidity, including cardiac, pulmonary, DM, stroke and so forth; cognitive status; functional status using level of nursing care Comorbidity, including cardiac, pulmonary disease, DM, malignancy, and so forth; functional status using level of mobility Comorbidity, including mainly cardiac and pulmonary disease
DM, diabetes mellitus; ILI, influenza-like illness. *In Newcastle-Ottawa quality assessment scale, any study is graded according to 3 categories: selection, comparability, and outcome. There are 4 items in selection, 1 item in comparability, and 3 items in outcome. A study can be awarded a maximum of 1 star (asterisk in this table) for each number item within the selection and outcome categories. A maximum of 2 stars can be given for comparability. In this review, 1 star would be given for comparability if age were compared between groups. Another star would be given if comorbidity and/or functional status were compared between groups in the study. The maximum possible score was 9 points. We took studies scoring
JAMDA journal homepage: www.jamda.com
Review
Effectiveness of Influenza Vaccination in Institutionalized Older Adults: A Systematic Review Tuen-Ching Chan MBBS, FHKAM a, b, *, Ivan Fan-Ngai Hung MD, FHKAM c, James Ka-Hay Luk MBBS, FHKAM a, Leung-Wing Chu MD, FHKAM b, Felix Hon-Wai Chan MB Bch, FHKAM a a b c
Department of Medicine and Geriatrics, Fung Yiu King Hospital, Hong Kong SAR, China Division of Geriatrics, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong SAR, China Division of Infectious Disease, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong SAR, China
a b s t r a c t Keywords: Influenza vaccination efficacy institutionalized older adults review meta-analysis
Introduction: Influenza infection is common among institutionalized older adults. Many nonrandomized observational studies on influenza vaccination suggested that it could reduce influenza-related hospitalizations and mortality in institutionalized older adults. Criticism regarding the effectiveness of influenza vaccine estimated by nonrandomized observational studies include the frailty selection bias and use of nonspecific outcome, such as all-cause mortality. Methods: We conducted a systematic review of studies of influenza vaccination in institutionalized older adults to determine the effects on clinical outcomes. We searched for studies from 3 databases from 1946 to June 2013 assessing effectiveness against influenza infection. We selected studies with good comparability between vaccine group and control group. We expressed vaccine effectiveness (VE) as a proportion, using the formula VE ¼ 1erelative risk or 1eodds ratio. We focused on the following outcomes: influenzalike illness (ILI), laboratory confirmed influenza, hospitalizations due to ILI, or pneumonia and death due to influenza or pneumonia. We did not include all-cause mortality. Results: Eleven studies that satisfied the inclusion criteria were identified, representing 11,262 institutionalized older adults. After meta-analysis, we found a significant reduction in pneumonia (VE: 37%, 95% confidence interval [CI]: 18%e53%, P ¼ .001) and death due to pneumonia or influenza (VE: 34%, CI: 10% e53%, P ¼ .01). There was no significant heterogeneity between studies. There was no significant publication bias. Conclusion: Influenza vaccination in institutionalized older adults could reduce pneumonia and death due to pneumonia or influenza. Influenza vaccination is recommended for institutionalized older adults. Ó 2014 - American Medical Directors Association, Inc. All rights reserved.
Influenza is a leading cause of infection death among older adults.1 The main strategy for prevention and control of seasonal and pandemic influenza for the past 50 years has been vaccination. The yearly influenza vaccination of at-risk individuals is a common practice, and adults 65 years and older (60 years in some countries) are suggested to have the vaccination.2,3 However, despite the widespread availability of influenza vaccination programs, vaccine coverage rates are generally unsatisfactory.4,5 This largely explains why influenza infections are still a major public health concern across the world.
The authors declare no conflicts of interest. * Address correspondence to Tuen-Ching Chan, MBBS, FHKAM, The University of Hong Kong, Division of Geriatrics, Department of Medicine, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong SAR, China. E-mail address: [email protected] (T.-C. Chan).
Institutionalized older adults have high infection-related hospitalization and mortality rates.6e9 High vaccine coverage in this population is very important. Most studies on influenza vaccination among older adults were nonrandomized observational studies. Previous systematic reviews on influenza vaccination suggested that influenza vaccination could reduce influenza-related hospitalizations and mortality in institutionalized older adults.10,11 However, the effectiveness of the influenza vaccine estimated by nonrandomized observational studies is frequently criticized for 2 major reasons: the frailty selection bias and use of nonspecific outcomes.12e14 Some physicians may feel that vaccine is futile in very frail older adults. A subset of undervaccinated frail elderly people could contribute a substantial proportion of adverse outcomes. If undervaccination were a direct consequence of the poor health status of those who are elderly, it would be the major source of bias. Many studies used nonspecific outcomes, especially all-cause mortality, that may contribute
1525-8610/$ - see front matter Ó 2014 - American Medical Directors Association, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jamda.2013.10.008
T.-C. Chan et al. / JAMDA 15 (2014) 226.e1e226.e6
to overestimation of vaccine efficacy.14 Therefore, we performed a systematic review and meta-analysis for effectiveness of influenza vaccination in institutionalized older adults with measures to minimize these confounding factors. Methodology We carried out a literature search using the following electronic databases: Medline 1946 to June 2013, Embase 1974 to June 2013, and the Cochrane Controlled Trials Register (The Cochrane Library 2013, Issue 4). We used the medical subject headings “influenza vaccine,” “influenza immunization,” “nursing home,” “institution,” “long-term care,” “elderly,” and “older adults.” Citations of relevant articles were reviewed. The Science Citation Index was used to trace any articles quoting a previous meta-analysis. Trial identification was based on PRISMA statement 2009.15 Studies with the following criteria were included: 1. Participants who were institutionalized adults aged 60 years or older. 2. Study design was a randomized controlled trial or observational study. 3. The outcome measures were all clinically relevant: clinically defined influenza-like illness (ILI), laboratory-confirmed influenza infection, pneumonia, hospitalization due to pneumonia or ILI, or death due to influenza or pneumonia. ILI was defined as fever together with respiratory and systemic signs and symptoms. Laboratory-confirmed influenza infection was defined as viral isolation with or without serological supporting evidence. Pneumonia was defined by clinical and radiological changes suggestive of pneumonia. We did not include all-cause mortality. 4. Data on outcomes of interest had to be reported in numerical format, including numerator and denominator data. Two observers independently extracted data and methodological details from the unmasked published reports. The following data were extracted: (1) description of the interventions, including the strain of the vaccine; (2) total number of subjects, and their age and sex; (3) follow-up duration; (4) study outcome definition and the number of outcome events corresponding to the requested definition in each group; (5) whether there was an outbreak in the nursing home during the study; and (6) whether the vaccine strain matched the circulating influenza strain. Discrepancies were documented and resolved by the consensus of both investigators. In a randomized controlled trial, Jadad’s16 system was used by the 2 reviewers to score each report. In Jadad’s16 system, 1 point is allocated for each of randomization, blinding, and description of withdrawals, and an additional point added for each of the well-described methods of randomization and blinding. Points were subtracted for clearly inadequate randomization or blinding. The maximum possible score was 5 points, and we took studies scoring less than 3 to be of low methodological quality. In nonrandomized observational studies, the Newcastle-Ottawa quality assessment scale was used by the 2 reviewers to score each study.17 In the Newcastle-Ottawa quality assessment scale, any study is graded according to 3 categories: selection, comparability, and outcome. There are 4 items in selection (representativeness of the exposed cohort/case control study; selection of the nonexposed cohort/case control study; ascertainment of exposures; and demonstration that outcome of interest was not present at start of study), 1 item in comparability (comparability of cohorts on the basis of the design or analysis), and 3 items in outcome (assessment of outcome; was followed long enough for outcomes to occur; and adequacy of follow-up of study). A study can
226.e2
be awarded a maximum of 1 star for each numbered item within the selection and outcome categories. A maximum of 2 stars can be given for comparability. In this review, 1 star would be given for comparability if age were compared between groups. Another star would be given if comorbidity and/or functional status were compared between groups in the study. The maximum possible score of Newcastle-Ottawa quality assessment scale was 9 points. We took studies scoring less than 8 to be of low methodological quality and only studies scoring 2 in the comparability category were included in the review. Election discrepancies were documented and resolved by the consensus of both observers. For each included study, if there were 2 or more nursing homes involved and outcomes of the 2 nursing homes were individually presented, data of each nursing home were individually analyzed during the meta-analysis (eg, Saito 2002a and Saito 2002b). Because in Saito 2002, there were 2 nursing homes in this cohort study. And the data of each individual nursing home was presented, hence the data of each nursing home was analyzed individually. Calculations were done with comprehensive meta-analysis software 2.0 (Biostat, Englewood, CA). The Mantel-Haenszel fixed effects model odds ratio (OR) was used to quantify the protective effect of the influenza vaccine. The heterogeneity between studies was assessed using the Cochrane Q statistical method, and a P value less than .10 was considered significant. When significant heterogeneity was found, a method based on a random effects model was used. Analyses were performed for different clinical outcomes. Within each clinical outcome, subgroup analyses were performed according to outbreak status in the nursing home and matching between vaccine strains and circulating influenza strain. There were 4 possible subgroup analyses for each clinical outcome: A, There was an outbreak, vaccine strain matched circulating influenza strain; B, There was an outbreak, vaccine strain unmatched with circulating influenza strain or unknown; C, There was no outbreak, vaccine strain matched circulating influenza strain; and D, There was no outbreak, vaccine strain unmatched with circulating influenza strain or unknown. Publication bias was assessed using funnel plot. We expressed vaccine effectiveness (VE) as a proportion, using the formula VE ¼ 1erelative risk or 1eOR.
Fig. 1. Flow diagram of study identification, based on the PRISMA statement.
226.e3
Table 1 Characteristics and Outcomes of Included Studies Study [Reference No.] Design
Sample Size
Outcome Measurement
Vaccine Matching
Quality Assessment Scales* Selection Comparability Outcome
Prospective cohort (1 y)
Currier 1988 [19]
Retrospective case-control for outbreak (6 mo)
Gross 1988 [20]
Prospective cohort study of outbreak (6 mo)
Horman 1986 [21]
Retrospective case control for outbreak (3 mo)
Leung 2007 [22]
Monto 2001 [23]
-Death due to pneumonia
Matched
****
**
***
-Clinically defined ILI -Clinically defined pneumonia
Not matched
***
**
***
-Laboratory-confirmed influenza -Clinically defined pneumonia
Matched
***
**
***
-Clinically defined ILI -Clinically defined pneumonia -Death due to ILI
Matched
****
**
***
3177 residents (2943 vaccinated and 234 controls) from 46 nursing homes
-Clinically defined ILI -Clinically defined pneumonia -Hospitalization due to ILI
Unknown
****
**
***
2351 residents (1728 vaccinated and 623 controls) from 26 nursing homes 128 residents (60 vaccinated and 68 controls) from a nursing home 1352 residents (678 vaccinated and 674 controls) from a nursing home 1629 residents (1074 vaccinated and 555 controls) from 20 nursing homes 84 residents (65 vaccinated and 19 controls) from a nursing home 97 residents (45 vaccinated and 52 controls) from a nursing home
-Clinically defined ILI -Clinically defined pneumonia -Death due to ILI -Clinically defined ILI -Hospitalization due to ILI -Death due to ILI -Clinically defined pneumonia -Death due to pneumonia
Matched
****
**
***
Matched
****
**
***
Comorbidity, including mainly cardiac and pulmonary disease
Matched in 1 y, **** other unmatched
**
***
-Clinically defined ILI
Matched
****
**
***
Comorbidity, including cardiac, pulmonary, stroke, DM, and so forth Comorbidity, including cardiac, pulmonary disease and stroke
-Clinically defined ILI -Death due to ILI
Matched
****
**
***
Comorbidity, including mainly cardiac and pulmonary disease
-Clinically defined ILI -Laboratory-confirmed influenza -Clinically defined pneumonia -Hospitalization due to ILI -Death due to ILI
Matched
****
**
***
Functional status using level of mobility
1859 residents (1214 vaccinated and 645 controls) from 58 nursing homes 121 residents (87 vaccinated and 34 controls) from a nursing home 305 residents (181 vaccinated and 124 controls) from a nursing home 159 residents (100 vaccinated and 59 controls) from a nursing home
Retrospective case control (1 y)
Prospective cohort study (6 mo)
Murayama 1999 [24] Retrospective case control for outbreak (6 mo) Saah 1986 [25]
Prospective cohort study (6 mo)
Saito 2002 [26]
Prospective cohort study (1 y)
Strassburg 1986 [27]
Retrospective case control study for outbreak (6 mo)
Taylor 1992 [28]
Retrospective case control study for outbreak (6 mo)
Comorbidity using comorbidity index; functional status using Barthel Index Comorbidity including cardiac and pulmonary disease; cognitive status Comorbidity, including cardiac, pulmonary, stroke, and so forth; medication Comorbidity, including cardiac, pulmonary, DM, stroke and so forth; cognitive status; functional status using level of nursing care Comorbidity, including cardiac, pulmonary disease, DM, malignancy, and so forth; functional status using level of mobility Comorbidity, including mainly cardiac and pulmonary disease
DM, diabetes mellitus; ILI, influenza-like illness. *In Newcastle-Ottawa quality assessment scale, any study is graded according to 3 categories: selection, comparability, and outcome. There are 4 items in selection, 1 item in comparability, and 3 items in outcome. A study can be awarded a maximum of 1 star (asterisk in this table) for each number item within the selection and outcome categories. A maximum of 2 stars can be given for comparability. In this review, 1 star would be given for comparability if age were compared between groups. Another star would be given if comorbidity and/or functional status were compared between groups in the study. The maximum possible score was 9 points. We took studies scoring
