Camouflaged sampling and contacting of people from administrative databases: reaching target patients without knowing who they are.

BACKGROUND: Methods are needed for using confidential data to select and reach patients without compromising their health privacy. METHODS: (a) SAMPLING: we created (1) an anonymous list of encrypted personal health numbers (EPHNs) of target patients (e.g., users of the medications of interest) and (2) an anonymous list of EPHNs of people randomly sampled from the general population of non-users. The two lists were merged and randomly ordered to make a camouflaged list. People on the second list were called camouflagers. Then EPHNs were matched to names and contact information, and the EPHNs were removed from the contact list. (b) Contacting: when we contacted patients by mail or telephone, we told them their names were selected from one of two lists, and their health status was unknown to us. We invited respondents to answer (1) a short survey for the general population or (2) a longer survey concerning the target condition. RESULTS: In five studies, the percentage of camouflagers--equal to one minus the positive predictive value of the camouflaged list--has varied from 10 to over 50%. This depended on the psychosocial sensitivity of the target medications or health conditions, the natural camouflaging by the population's heterogeneity or inaccuracy of the data on the target characteristic, the degree of stratification of the sample, and the cost of contacting. CONCLUSION: Camouflaging enables administrative data to be used for contacting target patients in selected populations while adhering to current data privacy laws and ethics principles.

The effects of coxib formulary restrictions on analgesic use and cost: regional evidence from Canada.

BACKGROUND: Public insurance plans for pharmaceuticals in Canada differ substantially across provinces in the conditions under which pharmaceuticals are reimbursed. Coxibs provide a good example. Québec had no restrictions on reimbursement for these drugs. Ontario required physicians to submit the clinical indications for their use on the prescription. British Columbia required physicians to seek and receive prior authorisation from the drug plan. OBJECTIVE: This study compares the effects of different reimbursement policies on coxib, non-selective non-steroidal anti-inflammatory drugs (nsNSAIDs), and gastro-protective agent (GPA) use and cost. STUDY DESIGN: Analysis of retrospective time series analysis of all NSAID and GPA administrative claims data from April 1997 through December 2002. SETTING: Administrative claims data from April 1997 through December 2002 for each of the publicly funded drug plans in Québec, Ontario, and British Columbia. In addition, we obtained data from BC PharmaNet, which records all dispensed prescriptions in British Columbia. PATIENTS OR OTHER PARTICIPANTS: Senior beneficiaries (>or= 65 years). MAIN OUTCOME MEASURE: We compared the projected total NSAID utilisation in the absence of coxib reimbursement restriction with actual utilisation by province and drug category. Projected utilisation was based on ARIMA modelling and reported as the number of defined daily doses (DDDs) per 100 senior (>or=65 years) beneficiaries/month. RESULTS: In Ontario, under its "limited use" policy, uptake and steady-state use of coxibs was similar to that in Québec, where there were no restrictions. In British Columbia, publicly funded use of coxibs was 6% of that in Ontario and Québec. Despite a shift to private reimbursement, total coxib use in BC was only 50% of use in Ontario and Québec. The use of all NSAIDS (nsNSAIDS plus coxibs) increased for all provincial drug plans except for BC. The increase and overall rate of total NSAID use was greatest in Ontario. Neither Ontario's nor BC's policies had an impact on use of nsNSAIDs or GPAs. CONCLUSION: Only BC's policy effectively limited publicly funded coxib use. However, there was substantial cost-shifting to out-of-pocket and third party insurance plans in BC.

Changes in rates of upper gastrointestinal hemorrhage after the introduction of cyclooxygenase-2 inhibitors in British Columbia and Ontario.

BACKGROUND: Population rates of upper gastrointestinal (GI) hemorrhage have been observed to increase with the introduction and rapid uptake of selective cyclooxygenase-2 (COX-2) inhibitors. Changes in COX-2 inhibitor use and upper GI bleeding rates in regions with relatively restrictive drug policies (e.g., British Columbia) have not been compared with changes in regions with relatively less restrictive drug policies (e.g., Ontario). METHODS: We collected administrative data for about 1.4 million people aged 66 years and older in British Columbia and Ontario for the period January 1996 to November 2002. We examined temporal changes in the prevalence of NSAID use and admissions to hospital because of upper GI hemorrhage in both provinces using cross-sectional time series analysis. RESULTS: During the period studied, the prevalence of NSAID use in British Columbia's population of older people increased by 25% (from 8.7% to 10.9%; p < 0.01), as compared with a 51% increase in Ontario (from 10.9% to 16.5%; p < 0.01). Hospital admissions because of upper GI hemorrhage increased significantly in Ontario by about 16% on average, or about 2 admissions per 10 000 elderly people, above expected values (p < 0.01). A similar increase was not observed in British Columbia. INTERPRETATION: More restrictive drug coverage policies, although limiting access to drugs and their potential benefits, may protect the population from adverse drug effects.

Using administrative healthcare data to recruit study subjects: experience with 'camouflaged sampling'.

PURPOSE: To recruit a sample of asthmatics heterogeneous for short-acting (SA) beta-agonist use while protecting their privacy, and to compare participants recruited via 'camouflaged sampling' to those recruited through media advertising. METHODS: Direct and indirect patient contact using camouflaged sampling was used to recruit a stratified random sample of asthmatics, identified based on their receipt of a prescription for a SA beta-agonist. Volunteers were recruited through media advertising. Recruitment rates were determined for both indirect and direct patient contact, and sampled participants were compared to volunteers recruited through media advertising for differences in SA beta-agonist use, demographic socioeconomic factors, and pulmonary function. RESULTS: 109 and 93 participants were recruited through camouflaged sampling and media advertising, respectively. Direct and indirect patient contact resulted in recruitment rates of 5 and 9%, respectively. Sampled participants were more heterogeneous for SA beta-agonist use, older, more likely to smoke or receive social assistance, and of lower socioeconomic status (SES) than volunteers. There was no difference in the association between SES and the magnitude of SA beta-agonist use between recruiting methods. CONCLUSIONS: Although recruitment rates were lower than anticipated, camouflaged sampling facilitated stratified sampling of a targeted population and resulted in a more heterogeneous sample while protecting patient privacy.

Client retention in the British Columbia Methadone Program, 1996-1999.

BACKGROUND: Methadone treatment for heroin addiction has been available for 40 years, but there is relatively little research on the effectiveness of Canadian programs. This paper describes one-year retention among the client cohorts entering the British Columbia Methadone Program during expansion between 1996 and 1999, and examines some factors previously shown to influence retention. METHODS: All methadone maintenance prescriptions dispensed to 1996-1999 program entrants were extracted from records of the BC Triplicate Prescription Program. Retention status and covariates were evaluated one year post-entry using logistic regression. Effects of retention status misclassification on time in the program were assessed with a Cox model for clients who received continuous daily dosing or short carries. RESULTS: Fifty-two percent of program entrants were still receiving methadone one year after entry; 24% had left the program at one year but later returned. Age at program entry and average daily dose of methadone were important predictors of continuation. In the logistic regression, only the 1999 year-of-entry cohort appeared to have a different retention trajectory. Year of entry is not a significant predictor of time in the program for those receiving daily or short carry doses only, and other results are consistent between models. INTERPRETATION: Retention rates in the BC Methadone Program are favourable and consistent with published rates. Program expansion does not reduce retention, once the effects of client age and dose are accounted for. Adequate daily dosing appears crucial to both initial retention and return to treatment.