Time-to-dormancy of patient health records at the University College Hospital, Ibadan, Nigeria

Résumé

Background: Experiences have shown that patient’s

record often become dormant due to cessation of

healthcare and retaining dormant records waste time

and resources for storage and may hinder retrieval

of active patients’ records. Knowledge of time-to

dormancy of these records is critical to formulation

of retention and disposition policies for patients’

records management. However, there is paucity of

information on time-to-dormancy of patient’s records

in Nigeria. This study aimed at identifying the most

suitable distribution for modelling time-to-dormancy,

estimate dormancy rate and associated factors of

records created at University College Hospital,

Ibadan, Nigeria.

Methods: Of the 84,613 patient records created from

1990-1994, a sample of 1537 was systematically

selected and reviewed. Information on patient’s

characteristics, including date of first and last visits

were extracted from each record using a data

extraction proforma. Data analyses were done using

descriptive statistics and Kaplan-Meier methods to

estimate time-to-dormancy and identify determinants

of time-to-dormancy. Records with single-entry

indicating patients with one contact were censored. Cox,

Exponential and Weibull survival models were fitted

to the time-to-dormancy data to test for best fit.

Results: Patient records that survived beyond the first

contact, indicated by two or more entries had a

Median Dormancy Time of 1.93 months with 95.0%

becoming dormant at about 151.89 (SE of 12.31)

months. Among the survival models tested the

Weibull yielded the best fit for patient’s records

dormancy time data. The hazard ratios for records

of admitted patients = 1.17 (95% CI, 1.53-5.75);

females = 1.10 (95%, CI, 0.95-1.25), treatment

outcome = 2.97 (95% CI, 1.53-5.75) were high,

(HR>1), indicating higher-risk of dormancy

compared to other patient characteristics examined.

Conclusion: Patient records followed Weibull

distribution with average time-to-dormancy of 2.8

years. Based on this, a records retention and

disposition policy of 13 years should be formulated.