Missing records in occupational disease epidemiology

Our plain-English summary of Missing records in occupational disease epidemiology, by P.E. Enterline, G.M. Marsh (Journal of Occupational Medicine., 1982).

The study

• Authors: P.E. Enterline, G.M. Marsh
• Journal: Journal of Occupational Medicine.
• Published: 1982;24(9):677-80
• Cited by: 7 articles, as counted by NIH iCite
• Read it: PubMed · Publisher
• In J-M's library: Johns-Manville's company medical library held a copy (J-M no. 001788).

In plain English

This paper described how the authors checked the completeness of worker records used in their mortality study of fibrous glass and mineral wool (man-made mineral fiber) workers at 15 plants. They compared a sample of 4,158 names from employers' quarterly social security tax reports with the record abstracts they had been given. Abstracts existed for 2,863 names. Of the 1,295 without abstracts, 621 had acceptable reasons, such as being female or salaried. Of the other 674, 393 had records in plant files that abstractors had missed and 281 had no record in the files. One plant's miss rate was 25%. Six plants were resurveyed; in one, more records were added than the original cohort held. In three resurveyed plants the missed records were clearly unusual in respiratory cancer deaths. For the whole cohort, adding missed records raised the respiratory cancer SMR from 102.4 to 104.7. A further plant was dropped because its records could not be verified. The authors concluded that cohorts should be verified and records carefully edited.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Methods report: verification of the completeness of the worker cohort in the authors' mortality study of man-made mineral fiber workers, by matching a sample of names from social security tax reports (Employers Quarterly Report on Earnings) against the cohort record abstracts
• Who was studied: Workers at 15 fibrous glass and mineral wool plant locations in the United States. A sample of 4,158 names and social security numbers was checked. A 16th plant could not be verified and was dropped.
• Exposure: Work in fibrous glass and mineral wool production (man-made mineral fibers). The paper concerns record completeness rather than exposure levels.
• Main results:
  • The total sample was 4,158 names. Abstracts were available for 2,863 and not available for 1,295. Of those 1,295, 621 had acceptable reasons. Of the other 674 (reason unacceptable), 393 (58.3%) had records in the plant files and 281 (41.7%) had no record in the files.
  • Plant 1 had a 25% miss rate with no reason found.
  • Six of the 15 plants were completely resurveyed. In one plant, more records were added than had been in the original cohort.
  • Nearly a third of the added (missed) cohort was found to be dead, compared with a quarter of the original cohort.
  • Among the six resurveyed plants, respiratory cancer SMRs for missed and original records were not much different in plants 1 and 8. In plants 3, 4 and 11, the missed records were clearly unusual in respiratory cancer mortality.
  • For the entire cohort, the SMR for respiratory cancer increased from 102.4 to 104.7 when all missed records were added.
  • For 53 people whose records were absent from company files, follow-up identified nine deaths. Where death certificates were found, cause of death was not unusual.
  • In the one plant that could not be verified (inactive records buried in a landfill), the unprocessed results looked like those expected if the plant had handled asbestos. The authors strongly suspected many records were missing, and the plant was dropped.
• The authors' conclusion: The authors concluded that because epidemiologic results can form the basis for regulatory decisions, cohorts should be verified and records carefully edited. They said they would be hesitant to put much weight on studies whose cohorts were not verified.
• Limitations the authors state:
  • The volume of names from the social security reports overwhelmed the authors' resources, so they had to check a sample rather than all names.
  • For people missing from plant files, the authors could not be certain they were truly eligible for the study.
  • The authors did not believe the absent-records groups were unusual, but said more follow-up was needed to be certain.
  • One plant could not be verified by this method and was excluded.

Topics

Medical terms in this study, explained in our Lexicon: Occupational Disease, Epidemiology.

Related studies

• Recovering missing mesothelioma deaths in death certificates using hospital records — V.S. Santana, E. Algranti, et al. (2018)
• Hospital records as a data source for occupational disease surveillance: a feasibility study — J. Balmes, D. Rempel, et al. (1992)
• Use of court records for supplementing occupational disease surveillance — E. Schwartz, P. Landrigan (1987)
• The potential impact of epidemiology on the prevention of occupational disease — D.H. Wegman (1992)
• Hospital records: an underutilized source of information regarding occupational diseases and exposures — G.M. Liss, R.A. Kusiak, et al. (1997)
• Extrapolation from occupational studies: a substitute for environmental epidemiology — P.E. Enterline (1981)

Browse more research in Journal of Occupational Medicine., or return to medical research.