Insurance Company Database Research Articles

The costs associated with the evaluation of rotator cuff tears before surgical repair

Patients undergoing rotator cuff repair typically have a diagnostic evaluation and trial of nonoperative therapy before surgery. Recent studies have evaluated the cost-effectiveness of surgery, but none have attempted to estimate the costs associated with the preoperative evaluation. This study used available data to examine major expenditures during the preoperative period. We conducted a search using an insurance company database to identify patients undergoing rotator cuff repair from 2004 to 2009. Patients were identified by the common Current Procedural Terminology codes for rotator cuff repair. The associated charge codes for the 90-day period before surgery were categorized as outpatient physician visits, diagnostic imaging studies, injections, physical therapy, laboratory and other preoperative studies, prior surgeries, and miscellaneous. The frequency of each code and the associated charges were noted. In total, 92,688 patients were identified in the study period. A total of $161,993,100 was charged during the preoperative period, for an average of $1,748 per patient. Diagnostic imaging charges totaled $104,510,646 (65%); injections, $5,145,227 (3%); outpatient visits, $29,723,751 (18%); physical therapy, $13,844,270 (8.5%); preoperative studies, $6,792,245 (4.2%); and miscellaneous, $1,164,688 (<1%). The costs for preoperative evaluation of rotator cuff tears are substantial, and the majority of the costs are associated with magnetic resonance imaging. To help reduce costs, future studies should attempt to identify the factors that predict which patients might not respond to nonoperative management and might benefit from early surgical intervention. In addition, magnetic resonance imaging should perhaps be reserved for patients in whom the diagnosis cannot be achieved by other modalities.

POPULATION COHORT STUDY TO EVALUATE THE HEALTH EFFECTS OF ENVIRONMENTAL AND OCCUPATIONAL EXPOSURES

Background: the area of Taranto in Southern Italy has been investigated in several environmental and epidemiological studies because of a large steel industry operating since the 1960’s. Excesses of mortality and cancer incidence were found in the local population, but the studies were mostly ecological, and did not consider the potential confounding effect of socioeconomic status (SES) or occupational exposures. Aims: we applied a cohort study design to evaluate cause-specific mortality and hospitalization in relation to PM10 exposure levels from the industry and occupational exposures. Methods: a cohort of residents (1998-2010) was enrolled. Industrial PM10 concentration (SPRAY dispersion model) was assigned to each participant on the basis of the geocoded addresses at the beginning of the follow-up. Individual occupational history was traced through the national insurance company (INPS) database. Several exposures were considered: districts of residence, SES, PM10 and occupational history (workers in steel, naval and mechanical construction factories vs others). Mortality/morbidity information were retrieved from Regional Health Databases. The associations between exposures and mortality/morbidity were estimated using multivariate Cox models. Results: 321,356 people were enrolled in the cohort. Higher mortality/morbidity risks were found for people with a low SES and resident in districts close to the industrial area. PM10 exposure (10 µg/m3) was associated with an increase risk of death from cardiac diseases (HR=1.04 95%CI 1.01-1.07), in particular acute coronary events (HR=1.09 95%CI 1.01-1.18) and respiratory hospitalizations (HR=1.08 95% CI 1.05-1.10). Workers in the steel plant (13,556 individuals) had an increased mortality for stomach, bladder and pleural cancer and increased hospitalization for cardiovascular and respiratory diseases. Conclusions: Exposure to PM10 from the industry and employment in the steel industry were associated with increased mortality/morbidity.

Incidence of Postoperative Infections Requiring Reoperation After Arthroscopic Knee Surgery

The purposes of this study were to determine the incidence of infection requiring reoperation after arthroscopic knee surgery during a 6-year period and to compare infection rates across different age groups, genders, geographic regions of the United States, and Current Procedural Terminology (CPT) codes through a retrospective review of a large insurance company database. A retrospective review of an insurance company database was performed for all knee arthroscopies performed in the United States from 2004 to 2009. The database was first queried for all knee arthroscopies, and the number of those cases requiring additional surgery for infection within 30 days was determined to calculate the incidence of infection. The incidence was stratified by gender, age group, region within the United States, and CPT code. A separate analysis for procedures using allografts was also performed. A total of 432,038 arthroscopic surgeries were performed, and the number of infections requiring drainage was 638, for an overall incidence of infection from 2004 to 2009 of 0.15%. Among adults, men were affected almost twice as often as women (P < .001), and among children, boys were affected almost 3 times as often as girls (P < .001). A decreasing incidence was noted in patients 60 years or older between 2004 and 2009 (P= .01). Overall, the incidence did not significantly vary by age, region, or CPT codes that specified the implantation of allograft tissue. Compared with diagnostic arthroscopy, the relative risk of infection was higher for CPT-29889 (posterior cruciate ligament reconstruction). The incidence of infection requiring reoperation after knee arthroscopy from 2004 to 2009 was 0.15%. The incidence was higher among male patients in both the adult and pediatric populations. The incidence of infection decreased from 2004 to 2009 in patients 60 years or older. Among adult patients, the incidence did not vary by age, by region, or by CPT codes that involved implantation of allografts. Level IV, cross-sectional study.

Impact of the European Randomized Study of Screening for Prostate Cancer (ERSPC) on prostate‐specific antigen (PSA) testing by Dutch general practitioners

To determine the impact of the European Randomized Study of Screening for Prostate Cancer (ERSPC) publication in 2009 on prostate-specific antigen (PSA) level testing by Dutch general practitioners (GPs) in men aged ≥40 years. Retrospective study with a Dutch insurance company database (containing PSA test claims) and a large district hospital-laboratory database (containing PSA-test results). The difference in primary PSA-testing rate as well as follow-up testing before and after the ERSPC was tested using the chi-square test with statistical significance at P < 0.05. Decline in PSA tests 4 months after ERSPC publication, especially for men aged ≥60 years. Primary testing as well as follow-up testing decreased, both for PSA levels of <4 ng/mL as well as for PSA levels of 4-10 ng/mL. Follow-up testing after a PSA level result of >10 ng/mL moderately increased (P = 0.171). Referral to a urologist after a PSA level result of >4 ng/mL decreased slightly after the ERSPC publication (P = 0.044). After the ERSPC publication primary PSA testing as well as follow-up testing decreased. Follow-up testing seemed not to be adequate after an abnormal PSA result. The reasons for this remain unclear.

The evaluation of antibiotic prescribing in general practice using a health insurance company database: penicillins use

General practitioners often prescribe antibiotics for diagnoses that do not require antibiotic treatment, partially because of the fear of treatment failure (and alibism). They are also strongly influenced by patients’ requests, and the empirical use of antibiotics is the main way of infection treatment in general practice in the Czech Republic. All this leads to non-adherence to the clinical recommendations for infections treatment. However, the regular monitoring of total antibiotic consumption does not reveal the critical areas of antibiotic prescription. Therefore, we used the database of a health insurance company to evaluate the antibiotic prescribing habits of general practitioners in an age-stratified population. In this short text, we report our findings on the prescribing of penicillins (Anatomical Therapeutic Chemical group J01C) as an example.

Rough sets and principal components analysis: A comparative study on customer database attributes selection

The databases of real world contain a huge volume of information; however, part of these data is not interesting for the knowledge extraction. So, data are preprocessed for reducing the amount of information and selecting more relevant attributes. This paper addresses a contrastive study between rough sets and principal components analysis on customer database attributes selection. The experiments were carried out using the insurance company database to evaluate a k-means clustering. The objective is to investigate the capacity of these techniques to improve the identification of customers segments in databases, what appears as an important tool to increase the effectiveness of business communication. Key words: Rough sets, principal components analysis, attributes selection, customer database.

Dental practitioners and ill health retirement: a qualitative investigation into the causes and effects

Data published in 1999 from the Government Actuary on the National Health Service Pension Scheme (1989-1994) have indicated that the frequency of ill health premature retirement (IHR) was four times more prevalent among dentists at age 42 years compared with doctors.Objective The aim of this project was to determine the factors that contributed to IHR in dental practitioners, and the effects of IHR on their lives. Semi-structured interviews were carried out during 2007. A topic list was developed, piloted and used to guide the interviewer. A purposive sampling technique was used to recruit the respondents from an insurance company database. A framework approach to data analysis was utilised. Twenty-three respondents were interviewed, 19 male and 4 female, aged between 39-59 years. Depression, stress and anxiety were reported by respondents to be major causes of their retirement, followed by musculoskeletal disease and premature disability caused by trauma. This is illustrated by the following: '...just went straight into general practice but with some regrets. Practice was so depressing' (GDS/NHS); 'I withdraw, I don't engage ... I found it more and more difficult and one morning... I collapsed in tears at the practice' (GDS/NHS). Dentists reported both negative and positive comments in relation to support received from their health insurance company. The main causes of IHR were depression, musculoskeletal disease and specific skin conditions. Respondents expressed concern regarding the level of support available to dentists in distress. Respondents to this study found that continuing to work had a positive impact on their health.

Quality of Care in Patients With Chronic Hepatitis C Virus Infection

Medicare has proposed quality-of-care indicators for chronic hepatitis C virus (HCV) infection. The extent to which these standards are met in practice is largely unknown. To evaluate the quality of health care that patients with HCV receive and the factors associated with receipt of quality care. Retrospective cohort study. Nationwide U.S. health insurance company research database. 10 385 patients with HCV enrolled in the database between 2003 and 2006. Patients were included if they were eligible for at least 1 quality indicator. Quality of HCV care received by patients, as measured by 7 explicit quality indicators included in Medicare's 2009 Physician Quality Reporting Initiative. Proportions of patients meeting quality indicators varied, ranging from 21.5% for vaccination to 79% for the HCV genotype testing indicator. Overall, 18.5% of patients (95% CI, 18% to 19%) received all recommended care. Older age and presence of comorbid conditions were associated with lower quality, whereas elevated liver enzyme levels, cirrhosis, and HIV infection were associated with higher quality. Patients who saw both generalists and specialists received the best care (odds ratio of receiving care for which a patient is eligible: specialists alone, 0.79 [CI, 0.66 to 0.95]; primary care physician alone, 0.44 [CI, 0.40 to 0.48]). The study had an observational retrospective design, used a convenience sample, and had no information on patient ethnicity. It may be that the indicators or the reporting of the indicators of HCV care--and not the care itself--is suboptimum. Health care quality, based on Medicare criteria, is suboptimum for HCV. Care that included both specialists and generalists is associated with the best quality. Our results support the development of specialist and primary care collaboration to improve the quality of HCV care. Saint Louis University Liver Center.

A Health Insurance Company–Initiated Multifaceted Intervention for Optimizing Acid-Suppressing Drug Prescriptions in Primary Care

To evaluate the effectiveness of a health insurance company-initiated intervention strategy aimed at optimizing acid-suppressing drug (ASD) prescriptions in primary care. In a cluster randomized controlled trial design, 112 primary care physician (PCP) peer review groups (993 PCPs) in the central region of the Netherlands were randomized. The PCPs in the intervention group received an ASD prescription optimization protocol, a list of their patients taking ASDs frequently on a long-term basis, and financial compensation for additional consultations with these patients. The PCPs in the control group did not receive any of these interventions. Prescription data on 23 433 patients were extracted from the database of the regional health insurance company. The main outcome measures were the proportion of patients who reduced ASD consumption by more than 50% and changes in annual volume and costs of ASD prescriptions. Differences in ASD reduction and in volume were analyzed applying multilevel regression analyses. At baseline, 2.4% of the patients (n = 967 506) of the participating practices used ASDs frequently on a long-term basis (>180 daily defined doses [DDDs] annually). During the 6-month intervention, 14.1% of the patients in the intervention group reduced ASD consumption compared with 13.7% in the control group (adjusted relative risk, 1.04; 95% confidence interval [CI], 0.97-1.11). Changes in intervention and control group in mean volume of ASD prescription per patient were similar (beta = 0.33 for DDD; 95% CI -3.00 to 3.60). A health insurance company-initiated multifaceted intervention, including practical tools and financial incentives, did not alter ASD prescription practice in primary care. More tailored interventions, including patient-targeted initiatives, are required to optimize ASD prescription.

Socio‐demographic factors influence chronic proton pump inhibitor use by a large population in the Netherlands

Chronic proton pump inhibitor (PPI) use is common in the Western world. Socio-economic status and socio-demographic factors have been shown to influence decisions related to prescribing of various drugs, but the influence of these factors on chronic PPI use is uncertain. To study the influence of SES and socio-demographic factors on chronic PPI use. Data were collected from a database of a Dutch health insurance company. Subjects having had at least one prescription for a PPI were identified and followed up for 6 months. Patients were then subdivided into chronic PPI users. Socio-demographic status was based on neighbourhood level of residence. Logistic regression was performed to determine socio-demographic factors associated with PPI use. A total of 2 001 787 insured individuals were included, 85 253 subjects were chronic users. Both low income (OR 1.55; CI 1.52-1.58) and low educational level (OR 1.33; CI 1.31-1.36) were associated with chronic PPI use. Other independent predictive variables included use of 10 or more concomitant medications (OR 5.33; CI 4.96-5.72) and the use of prokinetic drugs (OR 10.01; CI 9.22-10.88). Patients of a lower socio-demographic status are more likely to use PPIs on a chronic basis. The observed gradient in PPIs use may reflect differences in health, healthcare use or healthcare supply.

Colorado workers' compensation: medical vs chiropractic costs for the treatment of low back pain

Objective Low back disorders (LBDs) are the most common complaint among workers; therefore, many questions arise about cost-effective treatment approaches. This investigation evaluated the differences in cost-related factors among a population of patients selecting chiropractic vs allopathic care for the treatment of nonspecific LBDs. The study hypothesis was that chiropractic care would be more cost-effective or equivalent to allopathic care for the noncomplicated LBDs. Methods Cases were extracted from an insurance company database of patients reporting work-related low back injuries who were treated with either chiropractic or allopathic approaches. Cases were matched using the International Classification of Diseases, Ninth Revision, codes 722 (intervertebral disk disorders), 724 (other and unspecified disorders of the back), and 847 (sprains and strains of other and unspecified parts). The data set included 76 chiropractic cases and 2386 medical cases. Results The total amount paid by the insurance company was 1.7 times higher for patients treated by doctors of chiropractic (DCs) compared with those treated by medical doctors (MDs), and the cost of clinical treatment was 3.3 times higher for the DCs than MDs. Conclusion The cost for treatment by DCs was greater than that of MDs for similarly classified conditions affecting the low back. The amount paid by the insurance company was primarily related to the number of services given by each provider.

Preventing Falls in Physically Active Community-Dwelling Older People: A Comparison of Two Intervention Techniques

Background: Falls are common in physically active older people; however, most intervention studies have been targeted at frail older people. Objective: To evaluate the effectiveness of two intervention techniques for preventing falls in physically active community-dwelling older people. Methods: Two hundred and seventeen people (120 men and 97 women) aged between 70 and 90 years recruited from a health insurance company membership database were randomly allocated to receive either a psychomotor intervention focusing on body awareness, body experience and coordination, a fitness intervention focusing on functional skills, strength, endurance and flexibility, or no intervention (control group).The outcome measures wereprospective falls (number of fallers, number of multiple fallers, number of falls and falls rate) and measures of physical performance (Timed Up and Go Test, maximal step length, sit-to-stand time, normal and fast walking speed). Results: At the 4-month follow-up, significant improvements in the Timed Up and Go Test and sit-to-stand times were observed for both the psychomotor and fitness intervention groups. During the 12-month follow-up period, 39% of the participants fell at least once, and 19% fell on multiple occasions. The proportion of fallers in the fitness intervention group was 23% less than in the control group (RR = 0.77, 95% CI 0.60–0.97), but no significant reduction in falls was observed in the psychomotor intervention group. Conclusion: A fitness training program improves some aspects of physical performance and reduces falls by 23% in physically active older people.

An examination of periodontal treatment and per member per month (PMPM) medical costs in an insured population.

BackgroundChronic medical conditions have been associated with periodontal disease. This study examined if periodontal treatment can contribute to changes in overall risk and medical expenditures for three chronic conditions [Diabetes Mellitus (DM), Coronary Artery Disease (CAD), and Cerebrovascular Disease (CVD)].Methods116,306 enrollees participating in a preferred provider organization (PPO) insurance plan with continuous dental and medical coverage between January 1, 2001 and December 30, 2002, exhibiting one of three chronic conditions (DM, CAD, or CVD) were examined. This study was a population-based retrospective cohort study. Aggregate costs for medical services were used as a proxy for overall disease burden. The cost for medical care was measured in Per Member Per Month (PMPM) dollars by aggregating all medical expenditures by diagnoses that corresponded to the International Classification of Diseases, 9th Edition, (ICD-9) codebook. To control for differences in the overall disease burden of each group, a previously calculated retrospective risk score utilizing Symmetry Health Data Systems, Inc. Episode Risk Groups™ (ERGs) were utilized for DM, CAD or CVD diagnosis groups within distinct dental services groups including; periodontal treatment (periodontitis or gingivitis), dental maintenance services (DMS), other dental services, or to a no dental services group. The differences between group means were tested for statistical significance using log-transformed values of the individual total paid amounts.ResultsThe DM, CAD and CVD condition groups who received periodontitis treatment incurred significantly higher PMPM medical costs than enrollees who received gingivitis treatment, DMS, other dental services, or no dental services (p < .001). DM, CAD, and CVD condition groups who received periodontitis treatment had significantly lower retrospective risk scores (ERGs) than enrollees who received gingivitis treatment, DMS, other dental services, or no dental services (p < .001).ConclusionThis two-year retrospective examination of a large insurance company database revealed a possible association between periodontal treatment and PMPM medical costs. The findings suggest that periodontitis treatment (a proxy for the presence of periodontitis) has an impact on the PMPM medical costs for the three chronic conditions (DM, CAD, and CVD). Additional studies are indicated to examine if this relationship is maintained after adjusting for confounding factors such as smoking and SES.

Inzidenz proximaler Femurfrakturen in Deutschland

An increasing number of hip fractures were reported world wide during the last decades. Up to now incidence data are rare for Germany. The aim of this study is to determine actual person-related age- and sex-specific rates of hip fractures in Germany. The linkable and person-specific but non-identifying database of a German statutory health insurance company, Gmuender ErsatzKasse (GEK), was used. All subjects who were admitted to hospital with a primary diagnosis of hip fracture (ICD-10: S72.0-S72.2) in 2004 were identified. Incidence rate ratios (IRR), age- and sex-specific rates (per 1000 person-years) were calculated. Annual numbers of hip fractures for all persons covered by statutory health insurances (Gesetzliche Krankenversicherung- GKV) and the German population were estimated. The incidence of hip fracture increases exponentially with age. Hip fractures occur 129 times more often (IRR: 129.07; 95 % CI: 66.55 - 269.78) in women aged 90 - 94 years and 79 times more often (IRR: 78.97; 95 % CI: 30.02 - 185.87) in men aged 90 - 94 years compared to those aged 50 - 54 years. Approximately 101 037 (95 % CI: 91 779 - 110 294) persons are annually admitted to hospital in Germany due to this event. Hip fractures are common in the elderly and especially in women. Because of demographic changes and a possible secular trend an increasing number of hip fractures are to be expected in the future.

Utilization of fluoroquinolones andEscherichia coli resistance in urinary tract infection: inpatients and outpatients

The aim of this study was to evaluate the dependence of Escherichia coli resistance to fluoroquinolones on their use in the outpatients and inpatients in the Olomouc region of the Czech Republic. Data on inpatient antibiotic use were obtained from the database of the Department of Pharmacology and expressed as defined daily dose per 100 bed-days (DBD). Data on outpatient prescriptions were obtained from the database of General Health Insurance Company and expressed in defined daily doses per 1000 clients per day (DID). Escherichia coli strains were isolated from samples of urine of both community and hospitalized patients suffering from acute bacterial urinary tract infection, examined using aerobic cultivation, and determined by standard biochemical procedures. The utilization of fluoroquinolones in inpatients has significantly (p < 0.01) increased from 2.52 DBD in 1997 to 4.29 DBD in 2002. In outpatients, fluoroquinolone utilization has also increased significantly from 0.14 to 0.95 DID (p < 0.01). In the same period, 9192 E. coli strains were isolated from inpatients and outpatients suffering by urinary tract infections and tested for the susceptibility to fluoroquinolones. Resistance increased significantly (p < 0.01) both in the hospital (from 2 to 9%) and in the community (from 1 to 10%). The development of E. coli resistance to fluoroquinolones correlates significantly with their utilization both in the hospital (r = 0.944, p = 0.005) and in the community (r = 0.859, p = 0.029). Results of this study shows the impact of fluoroquinolone utilization on E. coli resistance and support the need of controlled use of these effective antibiotics.

Work injuries among Finnish farmers: A National Register Linkage study 1996–1997

Farming is one of the most injury-prone occupations in Finland as it is in other countries. Our goals were to describe work injuries of Finnish farmers and to compare occupational injury rates between various subgroups. A national cohort of 69,629 full-time farmers and their 11,657 compensated injuries were identified from an insurance company database. Cohort data were merged with a population census and farm register. Relative incidence rates were calculated using Poisson regression. Men had higher injury rates than women, except with regard to injuries caused by animals. Dairy and hog farming were the riskiest activities, and injury rates increased with the number of dairy cows. One-half of insured farmers in Finland are full-time farmers, which may have lead to underestimation of risk in Finnish injury statistics. Dairy farming is of particular concern because it is both common and has a high injury rate.

The impacts of health examinations and smoking on disease mortality risk in Japan: a longitudinal cohort of 720,611 Japanese life insured persons

To evaluate the impacts of health examinations (HE) and smoking on disease mortality risk in Japan. By using the large cohort database of a Japanese life insurance company, 720,611 subjects aged 20 to 80 years, who had contracted for life insurance between April 1, 1995 and March 31, 1998, were followed up until September 30, 1999. Cox's proportional hazard model was used to estimate age-adjusted relative risk (RR) for disease death. After adjusting for age, disease mortality in smokers was significantly higher than that in non-smokers (men, RR 1.51, 95% CI: 1.25-1.81; women, RR 1.54, 95% CI: 1.12-2.11). Meanwhile, disease mortality in HEees (those who had got HE within the past 2 years) was significantly lower than that in non-HEees (men, RR 0.70, 95% CI: 0.56-0.88; women, RR 0.71, 95% CI: 0.54-0.92). The magnitude of the impact of HE on disease mortality risk varied according to smoking status. Non-smokers showed a significantly lower risk associated with HE, whereas smokers did not. HE may allow an appreciable reduction in disease mortality, however, the reduction effect may be limited to non-smokers. Smoking cessation may be essential to improve the preventive effects of HE.

Epidemiological Evaluation of the Outcomes of Nonsurgical Root Canal Treatment in a Large Cohort of Insured Dental Patients

Selected outcomes following initial nonsurgical root canal treatment (NSRCT) procedures were retrospectively assessed using an insurance company database of 110,766 nonsurgical root canal procedures that were completed by endodontists and their referring general dentists. A subset of 44,613 cases, with a minimum required follow-up time of 2 yr, showed incidences of extraction, retreatment and periradicular surgery equal to 5.56%, 2.47%, and 1.41%, respectively. The incidence of subsequent extraction increased with patient age. Teeth that were not restored after root canal therapy were significantly more likely to undergo extraction than restored teeth. Although the practice pattern for endodontists consisted of a significantly higher proportion of molars (48% more; p < 0.001) and a smaller proportion of anterior teeth (43% less; p < 0.001) than general dentists, both groups of providers had comparable rates of untoward events. These data strongly support the hypothesis that the specialist practice provides similar rates of clinical success compared with other providers, even when treating significantly more complex NSRCT cases. Overall, 94.44% of nonsurgical root canal treated teeth remained functional over an average follow-up time of 3.5 yr. These results are an important indication of the benefits of endodontic treatment when provided in an integrated health care delivery system of endodontists and their referring general dentists.

CRM In A Real-world Insurance Company

CRM (Customer Relationship Management) is an iterative process that turns customer information into customer relationships. An important aspect regards customer Knowledge Discovery process which allows identification, segmentation and behavior prediction of the customer base. In this process data mining is probably the most known step. This paper presents a method for identifying the best prospects for a direct marketing campaign, consisting on the usage of two different classification techniques. The first one is based on demographic clustering and the other is based on neural network models, both of them are used against the prospect database. We apply this method to an insurance company database, using their customers' data in order to construct and verify the models. By comparing the results of the two analyses, we find that the two approaches almost always agree in the selection of the prospects, which are the best candidates for a particular policy. We show how the two techniques can be combined to improve the final result, by increasing the understanding on one side, and by decreasing the uncertainty and ranking the selected prospects, on the other side.

Validation of computerized Swedish dog and cat insurance data against veterinary practice records

Large computerized medical databases offer great potential for epidemiological research. However, data-quality issues must be addressed. This study evaluated the agreement between veterinary practice records and computerized insurance data in a large Swedish claims database. For the year 1995, the company insured over 320 000 dogs and cats. A total of 470 hard-copy records were sampled from claims for health care ( n=236) and life insurance ( n=234). Computerized insurance data for these claims were accessed and records from the attending veterinary practices were collected. For health and life claims, respectively, 79.2 and 72.8% of practice records were retrieved. Variables compared between the computer and practice records were breed, sex, date of birth and diagnosis for the claim. The degree of agreement was categorized as agreement, minor disagreement, major disagreement or data missing. Multivariable logistic regression was used to examine factors associated with errors. The observed agreement for breed and sex was excellent. There was 28.9 and 33.8% minor disagreement for the date of birth for health and life claims, respectively. This was mainly because, prior to 1993, the date of birth was coded as the first of January when more complete information was unavailable. Major disagreement (different year of birth) was low for both health and life claims. For health and life claims, the observed agreement for diagnosis was 84.0 and 84.9%, minor disagreement was 6.4 and 4.3%, and major disagreement 5.9 and 9.4%, respectively. Although there was no difference based on size of the veterinary clinic involved, there was a tendency for better agreement between the practice record and the computerized insurance data for claims from clinics with computerized practice records ( n=40) than for clinics with manual practice records ( n=286). Rates of discrepancy were affected by the clerk who processed the claims. Records processed by two of the 21 clerks had significantly more disagreements. Given the nature of the data, it was only possible to calculate a measure of observed agreement. In general, the agreement between data in the insurance-company database and from the practice records was excellent for demographic data such as breed and sex and fair for diagnostic information and date of birth. In general, the data are adequate to support ongoing research with due considerations of certain limitations.