Rapid ReportWho Doesn’t Authorize the Linking of Survey and Administrative Health Data? A General Population-based Investigation
Introduction
Linking survey and administrative data is becoming increasingly popular among epidemiologists and other health researchers because it brings about a confluence of virtues that each individual method cannot supply on its own and broadens the range of research questions that can be explored (1). Germane health data regulations such as the Health Insurance Portability and Accountability Act (HIPAA) of 1996 require individual authorization before such linking can take place. Individual authorization for linking is likely to be less than complete, raising concerns over what some have deemed “authorization bias” whereby those who authorize differ systematically from those who do not in terms of attributes important to health-related investigations (2). Although there is evidence that authorizers and nonauthorizers differ in terms of age, gender, income, and education 3, 4, 5, 6, 7, 8, 9, 10, past investigations are limited by their use of specialized patient populations and/or specific sociodemographic subgroups (11), potentially compromising the generalizability of their findings.
No one has attempted to characterize nonauthorizers in terms of their health and health care-seeking behavior in the context of an area-based population survey to our knowledge. The aim of this research is to describe who is willing to complete a survey but not willing to authorize use of their protected information in terms of their sociodemographic characteristics, medical and surgical diagnoses, comorbidities, and health care utilization by the use of linked survey and administrative health data from a general population-based study.
Section snippets
Methods
The data on authorization status (authorized vs. not) come from a sequential mixed mode, mail and telephone survey of noninstitutionalized residents of Olmsted County (home to Mayo Clinic), Minnesota, ages 18 and older as identified in a purchased list-based sample of residents. An experiment was embedded within this survey to determine the impact of including a HIPAA Authorization Form (HAF) on response rates in which half of the sample was randomized to be sent a HAF and the other half were
Results
A total of 1574 surveys or telephone interviews were completed. Of these, 1357 returned a completed HAF for an overall authorization rate of 86.2%. Comparing authorizers (n = 1357) to nonauthorizers (n = 217) demonstrated that subjects in the latter category were significantly younger, healthier (with respect to comorbidities), and less likely to have had three or more clinic office visits in 2005 and 2006 (Table 1). In a logistic model including all variables, age remained significant along
Discussion
Overall, 86.2% of those that completed a mailed survey or a telephone interview authorized linking to administrative health data, suggesting that some element of the HAF (evoked privacy concerns or signature requirement) was compelling some to not authorize, as observed by Singer 15, 16, 17. We found nonauthorizers to be significantly younger, healthier (with respect to comorbidities), and less likely to have had three or more recent clinic office visits than authorizers. However, once we
References (20)
- et al.
Potential effect of authorization bias on medical record research
Mayo Clin Proc
(1999) - et al.
Register-based study among employees showed small nonparticipation bias in health surveys and check-ups
J Clin Epidemiol
(2008) - et al.
Randomized trial showed an ‘embedded’ survey strategy optimized authorization rates compared to two ‘after survey’ strategies in veterans with PTSD
J Clin Epidemiol
(2010) - et al.
Randomized trial showed requesting medical records with a survey produced a more representative sample than requesting separately
J Clin Epidemiol
(2008) - et al.
Health services research using linked records: Who consents and what is the gain?
Aust N Z J Public Health
(2001) History of the Rochester Epidemiology Project
Mayo Clin Proc
(1996)- et al.
Linking survey data with administrative health information: Characteristics associated with consent from a neonatal intensive care unit follow-up study
Can J Public Health
(2005) - et al.
The impact of requiring patient authorization for use of data in medical records research
J Fam Pract
(1998) - et al.
Do characteristics of HIPAA consent forms affect the response rate?
Med Care
(2007) - et al.
Linking questionnaires to primary care records: Factors affecting consent in older people
J Epidemiol Community Health
(2005)
Cited by (0)
Supported by funds from the National Cancer Institute (R03 CA132974; PI: Beebe) and the Mayo Clinic Foundation for Education and Research. The study was made possible by the Rochester Epidemiology Project (R01 AG034676 from the National Institute on Aging; PI: Rocca).