Loss-to-follow-up may result in loss of statistical power by reduction of effective sample size; may introduce bias which compromises the study's validity [
[1]
]; result in different patient distribution between study arms [[2]
]; and underestimate mortality data [[3]
]. Reasons for loss-to-follow-up include deterioration in the medical condition, socioeconomic factors, and mortality [4
, 5
]. Tracking the survival status of patients who are lost-to-follow-up is limited by restricted access to death certificate data and patients moving. One strategy to obtain mortality data is to use online obituaries. Chronic disease clinics and studies have used online obituaries to track patients lost-to-follow-up [6
, 7
, 8
, 9
, 10
]. It is uncertain if this is appropriate for research. Our objectives were to evaluate the reliability and validity of online obituaries as a source of mortality data in two chronic diseases: systemic sclerosis (SSc) and idiopathic pulmonary arterial hypertension (IPAH).Subjects whose survival status was known were randomly selected from the Toronto Scleroderma Program and the University Health Network Pulmonary Hypertension Program. Five online obituary web sites were evaluated [
11
, Legacy.com. 2014. Available at: http://www.legacy.com/ns/. Accessed May 12, 2014.
12
, In Memorium. 2014. Death notices and obituaries in Canada. Available at: http://www.inmemoriam.ca/search/. Accessed May 14, 2014.
13
, Your Life Moments. 2014. Available at: http://www.yourlifemoments.ca/sitepages/static/obits/cityResults_44_1.asp. Accessed May 14, 2014.
14
, Life News. 2014. Available at: http://www.lifenews.ca/about. Accessed May 14, 2014.
15
]. Two investigators, blinded to survival status, independently entered the first and last name of each subject in each web site. If the subject was identified as deceased, other matching variables (date of birth, location, diagnosis) were used to verify the patient. Intrarater and interrater reliability was evaluated using the intraclass correlation coefficient (ICC). Pearson's correlation coefficient (r) was used to evaluate the correlation between the web site and actual survival status.Le Necrologue. 2014. Available at: http://www.lenecrologue.com/obituary/area/all/. Accessed May 16, 2014.
We studied 365 (219 SSc, 146 IPAH) subjects (Table 1). The ICC for the intrarater reliability was Legacy.com 0.95 (95% CI 0.93–0.96); Yourlifemoments.ca 0.96 (95% CI 0.95–0.97); Inmemoriam.ca 0.77 (95% CI 0.72–0.80), Lifenews.ca 0.75 (95% CI 0.71–0.80), and Lenecrologue.com 0.21 (95% CI 0.11–0.31). The ICC for interrater reliability using Legacy.com was 0.82 (95% CI 0.78–0.85). Because the intrarater reliability was lower for the other web sites, the ICC for interrater reliability for these web sites was not evaluated.
Table 1Patient characteristics
| Characteristics | Cohort, N = 365 | SSc, N = 219 | IPAH, N = 146 |
|---|---|---|---|
| Female sex | 273 (74.7%) | 171 (78.0%) | 102 (69.8%) |
| Deceased | 206 (56.4%) | 112 (51.1%) | 94 (64.3%) |
Abbreviations: SSc, systemic sclerosis; IPAH, idiopathic pulmonary arterial hypertension.
Legacy.com correctly identified the most deceased subjects (r = 0.36 [95% CI 0.27–0.45]). This positive correlation was similar across diseases (SSc r = 0.34 [95% CI 0.21–0.45]), IPAH (r = 0.41 [95% CI 0.26–0.53]) (Table 2).
Table 2Correlation of web site findings and known survival status
| Web site | Correlation coefficient, r (95% CI) |
|---|---|
| Legacy.com | |
| Total cohort | 0.36 (0.27, 0.45) |
| SSc | 0.34 (0.21, 0.45) |
| IPAH | 0.41 (0.26, 0.53) |
| Inmemoriam.ca | |
| Total cohort | 0.06 (−0.04, 0.16) |
| SSc | 0.08 (−0.05, 0.21) |
| IPAH | 0.01 (−0.15, 0.18) |
| Yourlifemoments.ca | |
| Total cohort | 0.09 (−0.01, 0.19) |
| SSc | 0.07 (−0.05, 0.21) |
| IPAH | 0.11 (−0.05, 0.27) |
| Lifenews.ca | |
| Total cohort | 0.08 (−0.02, 0.18) |
| SSc | 0.02 (−0.11, 0.15) |
| IPAH | 0.16 (−0.01, 0.32) |
| Lenecrologue.com | |
| Total cohort | 0.12 (−0.02, 0.18) |
| SSc | 0.09 (−0.04, 0.22) |
| IPAH | 0.19 (0.03, 0.34) |
Abbreviations: CI, confidence interval; SSc, systemic sclerosis; IPAH, idiopathic pulmonary arterial hypertension.
We found that selected obituary web sites were reliable and valid sources of survival data. Traditional sources of survival data include clinic charts, hospital medical records, or death certificates. However, these sources have critical limitations [
[16]
]. Outpatient charts are dependent on notification that the patient is deceased or has moved. Hospital records are only useful if the terminal event occurred as an inpatient. Most investigators do not have access to death certificates. Furthermore, death certificates have been shown to be an inaccurate source of cause of death data as individuals unaware of the full medical details, or insufficiently trained in accurately indicating the required information, frequently complete them [16
, 17
]. The inaccuracy of death certificates is particularly true in older patients that have a greater number of comorbidities [[18]
].In summary, use of selected online obituaries is a reliable and valid method to collect mortality data. This approach may be used to track the vital status of some patients who are lost-to-follow-up.
References
- Loss to follow-up in cohort studies: how much is too much?.Eur J Epidemiol. 2004; 19: 751-760
- Underestimation of mortality following lung volume reduction surgery resulting from incomplete follow-up.Chest. 2001; 119: 1056-1060
- Potential impact on estimated treatment effects of information lost to follow-up in randomised controlled trials (LOST-IT): systematic review.BMJ. 2012; 344: e2809
- Understanding reasons for and outcomes of patients lost to follow-up in antiretroviral therapy programs in Africa through a sampling-based approach.J Acquir Immune Defic Syndr. 2010; 53: 405-411
- Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment programmes in sub-Saharan Africa.PLoS Med. 2011; 8: e1000390
- Study design, precision, and validity in observational studies.J Palliat Med. 2009; 12: 77-82
- The effect of male sex on survival in systemic sclerosis.J Rheumatol. 2014; 41: 2193-2200
- Warfarin in systemic sclerosis-associated and idiopathic pulmonary arterial hypertension. A Bayesian approach to evaluating treatment for uncommon disease.J Rheumatol. 2012; 39: 276-285
- Survival in rheumatoid arthritis-associated pulmonary arterial hypertension compared with idiopathic pulmonary arterial hypertension.Respirology. 2015; 20: 481-487
- Sex disparities in systemic sclerosis-associated pulmonary arterial hypertension: a cohort study.Arthritis Res Ther. 2016; 18: 30
Legacy.com. 2014. Available at: http://www.legacy.com/ns/. Accessed May 12, 2014.
In Memorium. 2014. Death notices and obituaries in Canada. Available at: http://www.inmemoriam.ca/search/. Accessed May 14, 2014.
Your Life Moments. 2014. Available at: http://www.yourlifemoments.ca/sitepages/static/obits/cityResults_44_1.asp. Accessed May 14, 2014.
Life News. 2014. Available at: http://www.lifenews.ca/about. Accessed May 14, 2014.
Le Necrologue. 2014. Available at: http://www.lenecrologue.com/obituary/area/all/. Accessed May 16, 2014.
- “Inaccuracy” in death certification–where are we now?.J Public Health Med. 1996; 18: 59-66
- Death certification reform: a case study on the potential impact on mortality statistics England and Wales. Statistical Bulletin. Experimental Statistics.Office for National Statistics, UK2012: 1-17
- Mortality in Rheumatoid Arthritis (RA): factors associated with recording RA on death certificates.BMC Musculoskelet Disord. 2015; 16: 277
Article info
Publication history
Published online: June 29, 2016
Footnotes
Funding: S.R.J. is supported by a Canadian Institutes of Health Research Clinician Scientist Award, the Oscar and Eleanor Markovitz Fund for Scleroderma Research, and the Freda Fejer Fund of the Arthritis Research Foundation.
Identification
Copyright
© 2016 Elsevier Inc. All rights reserved.
