Predicting the cost of illness: a comparison of alternative models applied to stroke

Does the Addition of Either a Lateral or Posterior Interbody Device to Posterior Instrumented Lumbar Fusion Decrease Cost Over a 6-Year Period?

STUDY DESIGN

Retrospective case-control study.

OBJECTIVES

Few studies have compared the costs of single-level (1) posterior instrumented fusion alone (PSF), (2) posterior interbody fusion with PSF (PLIF), and (3) lateral interbody fusion with PSF (circumferential LLIF). The purpose of this study was to compare costs associated with these procedures.

METHODS

Charts were reviewed and patients followed-up with a telephone questionnaire. Medicare reimbursement data was used for cost estimation from the payer's perspective. Multivariate survival analysis was performed to assess time to elevated resource use (greater than 90% of study patients or $68 672).

RESULTS

A total of 337 patients (PSF, 45; PLIF, 222; circumferential LLIF, 70) were included (63% follow-up at 6 years). PSF and circumferential LLIF patients were 3 times more likely to reach the cutoff value compared with PLIF patients (P = .017).

CONCLUSIONS

Circumferential LLIF and PSF patients were more likely to have higher resource use than PLIF patients and thus incur greater costs at 6-year follow-up.

Variable selection for random effects two-part models

Random effects two-part models have been applied to longitudinal studies for zero-inflated (or semi-continuous) data, characterized by a large portion of zero values and continuous non-zero (positive) values. Examples include monthly medical costs, daily alcohol drinks, relative abundance of microbiome, etc. With the advance of information technology for data collection and storage, the number of variables available to researchers can be rather large in such studies. To avoid curse of dimensionality and facilitate decision making, it is critically important to select covariates that are truly related to the outcome. However, owing to its intricate nature, there is not yet a satisfactory variable selection method available for such sophisticated models. In this paper, we seek a feasible way of conducting variable selection for random effects two-part models on the basis of the recently proposed "minimum information criterion" (MIC) method. We demonstrate that the MIC formulation leads to a reasonable formulation of sparse estimation, which can be conveniently solved with SAS Proc NLMIXED. The performance of our approach is evaluated through simulation, and an application to a longitudinal alcohol dependence study is provided.

Understanding Dynamic Status Change of Hospital Stay and Cost Accumulation via Combining Continuous and Finitely Jumped Processes

The Coxian phase-type models and the joint models of longitudinal and event time have been extensively used in the studies of medical outcome data. Coxian phase-type models have the finite-jump property while the joint models usually assume a continuous variation. The gap between continuity and discreteness makes the two models rarely used together. In this paper, a partition-based approach is proposed to jointly model the charge accumulation process and the time to discharge. The key construction of our new approach is a set of partition cells with their boundaries determined by a family of differential equations. Using the cells, our new approach makes it possible to incorporate finite jumps induced by a Coxian phase-type model into the charge accumulation process, therefore taking advantage of both the Coxian phase-type models and joint models. As a benefit, a couple of measures of the "cost" of staying in each medical stage (identified with phases of a Coxian phase-type model) are derived, which cannot be approached without considering the joint models and the Coxian phase-type models together. A two-step procedure is provided to generate consistent estimation of model parameters, which is applied to a subsample drawn from a well-known medical cost database.

Costs associated with Barrett's esophagus screening in the community: an economic analysis of a prospective randomized controlled trial of sedated versus hospital unsedated versus mobile community unsedated endoscopy

BACKGROUND AND AIMS

Data on the economic impact associated with screening for Barrett's esophagus (BE) are limited. As part of a comparative effectiveness randomized trial of unsedated transnasal endoscopy (uTNE) and sedated EGD (sEGD), we assessed costs associated with BE screening.

METHODS

Patients were randomly allocated to 3 techniques: sEGD or uTNE in a hospital setting (huTNE) versus uTNE in a mobile research van (muTNE). Patients were called 1 and 30 days after screening to assess loss of work (because of the screening procedure) and medical care sought after procedure. Direct medical costs were extracted from billing claims databases. Indirect costs (loss of work for subject and caregiver) were estimated using patient reported data. Statistical analyses including multivariable analysis accounting for comorbidities were conducted to compare costs.

RESULTS

Two hundred nine patients were screened (61 sEGD, 72 huTNE, and 76 muTNE). Thirty-day direct medical costs and indirect costs were significantly higher in the sEGD than the huTNE and muTNE groups. Total costs (direct medical + indirect costs) were also significantly higher in the sEGD than in the uTNE group. The muTNE group had significantly lower costs than the huTNE group. Adjustment for age, sex, and comorbidities on multivariable analysis did not change this conclusion.

CONCLUSIONS

Short-term direct, indirect, and total costs of screening are significantly lower with uTNE compared with sEGD. Mobile uTNE costs were lower than huTNE costs, raising the possibility of mobile screening as a novel method of screening for BE and esophageal adenocarcinoma.

Direct medical costs of accidental falls for adults with transfemoral amputations

BACKGROUND

Active individuals with transfemoral amputations are provided a microprocessor-controlled knee with the belief that the prosthesis reduces their risk of falling. However, these prostheses are expensive and the cost-effectiveness is unknown with regard to falls in the transfemoral amputation population. The direct medical costs of falls in adults with transfemoral amputations need to be determined in order to assess the incremental costs and benefits of microprocessor-controlled prosthetic knees.

OBJECTIVE

We describe the direct medical costs of falls in adults with a transfemoral amputation.

STUDY DESIGN

This is a retrospective, population-based, cohort study of adults who underwent transfemoral amputations between 2000 and 2014.

METHODS

A Bayesian structural time series approach was used to estimate cost differences between fallers and non-fallers.

RESULTS

The mean 6-month direct medical costs of falls for six hospitalized adults with transfemoral amputations was US$25,652 (US$10,468, US$38,872). The mean costs for the 10 adults admitted to the emergency department was US$18,091 (US$-7,820, US$57,368).

CONCLUSION

Falls are expensive in adults with transfemoral amputations. The 6-month costs of falls resulting in hospitalization are similar to those reported in the elderly population who are also at an increased risk of falling. Clinical relevance Estimates of fall costs in adults with transfemoral amputations can provide policy makers with additional insight when determining whether or not to cover a prescription for microprocessor-controlled prosthetic knees.

Developing a standardized healthcare cost data warehouse

Background

Research addressing value in healthcare requires a measure of cost. While there are many sources and types of cost data, each has strengths and weaknesses. Many researchers appear to create study-specific cost datasets, but the explanations of their costing methodologies are not always clear, causing their results to be difficult to interpret. Our solution, described in this paper, was to use widely accepted costing methodologies to create a service-level, standardized healthcare cost data warehouse from an institutional perspective that includes all professional and hospital-billed services for our patients.

Methods

The warehouse is based on a National Institutes of Research–funded research infrastructure containing the linked health records and medical care administrative data of two healthcare providers and their affiliated hospitals. Since all patients are identified in the data warehouse, their costs can be linked to other systems and databases, such as electronic health records, tumor registries, and disease or treatment registries.

Results

We describe the two institutions’ administrative source data; the reference files, which include Medicare fee schedules and cost reports; the process of creating standardized costs; and the warehouse structure. The costing algorithm can create inflation-adjusted standardized costs at the service line level for defined study cohorts on request.

Conclusion

The resulting standardized costs contained in the data warehouse can be used to create detailed, bottom-up analyses of professional and facility costs of procedures, medical conditions, and patient care cycles without revealing business-sensitive information.

After its creation, a standardized cost data warehouse is relatively easy to maintain and can be expanded to include data from other providers. Individual investigators who may not have sufficient knowledge about administrative data do not have to try to create their own standardized costs on a project-by-project basis because our data warehouse generates standardized costs for defined cohorts upon request.

Electronic supplementary material

The online version of this article (doi:10.1186/s12913-017-2327-8) contains supplementary material, which is available to authorized users.

Hospital Use by Persons With HIV in the 21st Century: A 5-State Study

OBJECTIVE

The purpose of this study is to determine whether reductions in hospital utilization observed immediately after the availability of highly active antiretroviral therapy (between 1995 and 2000) have persisted into the 21st century.

DATA SOURCES

Data on all human immunodeficiency virus (HIV)-related hospital admissions in 5 states (California, Florida, New Jersey, New York, and South Carolina) in 2000, 2005, 2010, and 2013 were obtained from the State Inpatient Database, which is administered by the Agency for Healthcare Research and Quality. In addition, data on the number of persons living with HIV were obtained from the Centers for Disease Control and Prevention and from the California Department of Public Health.

STUDY DESIGN

This study compares the average number of hospitalizations per person living with HIV in each of the 5 states as well as the average cost for hospital care per person with HIV in 2000, 2005, 2010, and 2013.

RESULTS

The total number of hospitalizations by persons with HIV in the 5 study states fell by one third between 2000 and 2013 even though the number of persons living with HIV increased by >50%.

CONCLUSIONS

Persons with HIV disease were 64% less likely to be hospitalized in 2013 than they were in 2000. In addition, the probability of a person with HIV being hospitalized fell 44% between 2000 and 2010 and 29% between 2010 and 2013.

Real-world costs of ischemic stroke by discharge status

OBJECTIVE

The objective of this study was to estimate the acute healthcare costs of ischemic stroke during hospitalization and the quarterly all-cause healthcare costs for the first year after discharge by discharge status.

METHODS

Adult patients with a hospitalization with a diagnosis of ischemic stroke (ICD-9-CM: 434.xx or 436.xx) between 1 January 2006 and 31 March 2015 were identified from a large US commercial claims database. Patients were classified into three cohorts based on their discharge status from the first stroke hospitalization, i.e. dead at discharge, discharged with disability, or discharged without disability. Third-party (medical and pharmacy) and out-of-pocket costs were adjusted to 2015 USD.

RESULTS

A total of 7919 patients dead at discharge, 45,695 patients discharged with disability, and 153,778 patients discharged without disability were included in this analysis. The overall average age was 59.7 years and 52.3% were male. During hospitalization, mean total costs (third-party and out-of-pocket) were $68,370 for patients dead at discharge, $73,903 for patients discharged with disability, and $24,448 for patients discharged without disability (p < .001 for each pairwise comparison); mean third-party costs were $63,605 for patients dead at discharge, $67,861 for patients discharged with disability and $19,267 for patients discharged without disability (p < .001 for each pairwise comparison). During the first year after discharge, mean total costs for patients discharged with disability vs. without disability were $46,850 vs. $30,132 (p < .001). Mean third-party costs for patients discharged with disability vs. without disability were $19,116 vs. $10,976 during the first quarter after discharge, $10,236 vs. $6926 during the second quarter, $8241 vs. $5810 during the third quarter, and $6875 vs. $5292 during the fourth quarter (p < .001 for each quarter).

CONCLUSION

The results demonstrated the high economic burden of ischemic stroke, especially among patients discharged with disability with the highest costs incurred during the inpatient stays.

Multiple chronic conditions and healthcare costs among adults

The prevalence of multiple chronic conditions (MCC) is increasing among individuals of all ages. MCC are associated with poor health outcomes. The presence of MCC has profound healthcare utilization and cost implications for public and private insurance payers, individuals, and families. Investigators have used a variety of definitions for MCC to evaluate costs associated with MCC. The objective of this article is to examine the current literature in estimating excess costs associated with MCC among adults. The discussion highlights some of the theoretical and technical merits of various MCC definitions and models used to estimate the excess costs associated with MCC.

Is the choice of the statistical model relevant in the cost estimation of patients with chronic diseases? An empirical approach by the Piedmont Diabetes Registry

BACKGROUND

Chronic diseases impose large economic burdens. Cost analysis is not straightforward, particularly when the goal is to relate costs to specific patterns of covariates, and to compare costs between diseased and healthy populations. Using different statistical methods this study describes the impact on results and conclusions of analyzing health care costs in a population with diabetes.

METHODS

Direct health care costs of people living in Turin were estimated from administrative databases of the Regional Health System. Patients with diabetes were identified through the Piedmont Diabetes Registry. The effect of diabetes on mean annual expenditure was analyzed using the following multivariable models: 1) an ordinary least squares regression (OLS); 2) a lognormal linear regression model; 3) a generalized linear model (GLM) with gamma distribution. Presence of zero cost observation was handled by means of a two part model.

RESULTS

The OLS provides the effect of covariates in terms of absolute additive costs due to the presence of diabetes (€ 1,832). Lognormal and GLM provide relative estimates of the effect: the cost for diabetes would be six fold that for non diabetes patients calculated with the lognormal. The same data give a 2.6-fold increase if calculated with the GLM. Different methods provide quite different estimated costs for patients with and without diabetes, and different costs ratios between them, ranging from 3.2 to 5.6.

CONCLUSIONS

Costs estimates of a chronic disease vary considerably depending on the statistical method employed; therefore a careful choice of methods to analyze data is required before inferring results.

International Comparison of Poststroke Resource Use: A Longitudinal Analysis in Europe

BACKGROUND

Long-term costs often represent a large proportion of the total costs induced by stroke, but data on long-term poststroke resource use are sparse, especially regarding the trajectory of costs by severity. We used a multinational longitudinal survey to estimate patterns of poststroke resource use by degree of functional disability and to compare resource use between regions.

METHODS

The Survey of Health, Ageing and Retirement in Europe (SHARE) is a multinational database of adults 50 years and older, which includes demographic information about respondents, age when stroke first occurred, current activity of daily living (ADL) limitations, and health care resource use in the year before interview. We modeled resource use with a 2-part regression for number of hospital days, home nursing hours, and paid and unpaid home caregiving hours.

RESULTS

After accounting for time since stroke, number of strokes and comorbidities, age, gender, and European regions, we found that poststroke resource use was strongly associated with ADL limitations. The duration since the stroke event was significantly associated only with inpatient care, and informal help showed significant regional heterogeneity across all ADL limitation levels.

CONCLUSIONS

Poststroke physical deficits appear to be a strong driver of long-term resource utilization; treatments that decrease such deficits offer substantial potential for downline cost savings. Analyzing internationally comparable panel data, such as SHARE, provide valuable insight into long-term cost of stroke. More comprehensive international comparisons will require registries with follow-up, particularly for informal and formal home-based care.

Smoking status and health care costs in the perioperative period: a population-based study

IMPORTANCE

Cigarette smoking adds an estimated $100 billion in annual incremental direct health care costs nationwide. Cigarette smoking increases complication risk in surgical patients, but the potential effects of smoking status on perioperative health care costs are unclear.

OBJECTIVE

To test the hypothesis that current and former smoking at the time of admission for inpatient surgery, compared with never smoking, are independently associated with higher incremental health care costs for the surgical episode and the first year after hospital discharge.

DESIGN, SETTING, AND PARTICIPANTS

This population-based, propensity-matched cohort study, with cohort membership based on smoking status (current smokers, former smokers, and never smokers) was performed at Mayo Clinic in Rochester (a tertiary care center) and included patients at least 18 years old who lived in Olmsted County, Minnesota, for at least 1 year before and after the index surgery.

EXPOSURE

Undergoing an inpatient surgical procedure at Mayo Clinic hospitals between April 1, 2008, and December 31, 2009.

MAIN OUTCOMES AND MEASURES

Total costs during the index surgical episode and 1 year after hospital discharge, with the latter standardized as costs per month. Costs were measured using the Olmsted County Healthcare Expenditure and Utilization Database, a claims-based database including information on medical resource use, associated charges, and estimated economic costs for patients receiving care at the 2 medical groups (Mayo Clinic and Olmsted Medical Center) that provide most medical services within Olmsted County, Minnesota.

RESULTS

Propensity matching resulted in 678 matched pairs in the current vs never smoker grouping and 945 pairs in the former vs never smoker grouping. Compared with never smokers, adjusted costs for the index hospitalization did not differ significantly for current or former smokers. However, the adjusted costs in the year after hospitalization were significantly higher for current and former smokers based on regression analysis (predicted monthly difference of $400 [95% CI, $131-$669] and $273 [95% CI, $56-$490] for current and former smokers, respectively).

CONCLUSIONS AND RELEVANCE

Compared with never smokers, health care costs during the first year after hospital discharge for an inpatient surgical procedure are higher in both former and current smokers, although the cost of the index hospitalization is not affected by smoking status.

Cost differences among robotic, vaginal, and abdominal hysterectomy

OBJECTIVE

To compare the costs of vaginal and abdominal hysterectomy with robotically assisted hysterectomy.

METHODS

We identified all cases of robotically assisted hysterectomy, with or without bilateral salpingo-oophorectomy, treated at the Mayo Clinic (Rochester, Minnesota) from January 1, 2007, through December 31, 2009. Cases were propensity score-matched (one-to-one) to cases of vaginal and abdominal hysterectomy, selected randomly from January 1, 2004, through December 31, 2006 (before acquisition of the robotic surgical system). All billed costs were abstracted through the sixth postoperative week from the Olmsted County Healthcare Expenditure and Utilization Database and compared between cohorts with a generalized linear modeling framework. Predicted costs were estimated with the recycled predictions method. Costs of operative complications also were estimated.

RESULTS

The total number of abdominal hysterectomies collected for comparison was 234 and the total number of vaginal hysterectomies was 212. Predicted mean cost of robotically assisted hysterectomy was $2,253 more than that of vaginal hysterectomy ($13,619 compared with $11,366; P<.001), although costs of complications were not significantly different. The predicted mean costs of robotically assisted compared with abdominal hysterectomy were similar ($14,679 compared with $15,588; P=.35). The costs of complications were not significantly different.

CONCLUSIONS

Overall, vaginal hysterectomy was less costly than robotically assisted hysterectomy. Abdominal hysterectomy and robotically assisted hysterectomy had similar costs.

LEVEL OF EVIDENCE

II.

Health care costs for state transition models in prostate cancer

OBJECTIVE

To obtain estimates of direct health care costs for prostate cancer (PC) from diagnosis to death to inform state transition models.

METHODS

A stratified random sample of PC patients residing in 3 geographically diverse regions of Ontario, Canada, and diagnosed in 1993-1994, 1997-1998, and 2001-2002, was selected from the Ontario Cancer Registry. We retrieved patients' pathology reports to identify referring physicians and contacted surviving patients and next of kin of deceased patients for informed consent. We reviewed clinic charts to obtain data required to allocate each patient's observation time to 11 PC-specific health states. We linked these data to health care administrative databases to calculate resource use and costs (Canadian dollars, 2008) per health state. A multivariable mixed-effects model determined predictors of costs.

RESULTS

The final sample numbered 829 patients. In the regression model, total direct costs increased with age, comorbidity, and Gleason score (all P < 0.0001). Radical prostatectomy was the most costly primary treatment health state ($4676 per 100 days). Radical prostatectomy, hormone-refractory metastatic disease ($6398 per 100 days), and final (predeath) ($13,739 per 100 days) health states were significantly more costly (P < 0.05) than nontreated nonmetastatic PC ($3440 per 100 days), whereas the postprostatectomy ($732 per 100 days) and postradiation ($1556 per 100 days) states cost significantly less (P < 0.0001).

CONCLUSIONS

This study used an innovative but labor-intensive approach linking chart and administrative data to estimate health care costs. Researchers should weigh the potential benefits of this method against what is involved in implementation. Modifications in methodology may achieve similar gains with less outlay in individual studies. However, we believe that this is a promising approach for researchers wishing to advance the quality of costing in state transition modeling.

Impact of a palliative care consult service

Established hospital palliative care consult services (PCCS) have been associated with reduced costs and length of stay, decreased symptom burden, and increased satisfaction with care. Using a retrospective case-control design, we analyzed administrative data of patients seen by PCCS while hospitalized at the Rochester, Minnesota Mayo Clinic hospitals from 2003 to 2008. The PCCS patients were matched to 3:1. A total of 1477 patients seen by the PCCS were matched with 4431 patients not seen. Costs for patients seen and discharged alive were US $35,449 (95% confidence interval [CI] US $34,157-US $36,686) compared to US $37,447 (95% CI US $36,734-US $38,126), without PCCS consultation. Costs for PCCS patients that died during hospitalization were US $54,940 (95% CI US $51,483-US $58,576) and non-PCCS patients were US $79,660 (95% CI US $76,614-US $83,398).

Predictors and costs of surgical site infections in patients with endometrial cancer

OBJECTIVE

Technological advances in surgical management of endometrial cancer (EC) may allow for novel risk modification in surgical site infection (SSI).

METHODS

Perioperative variables were abstracted from EC cases surgically staged between January 1, 1999, and December 31, 2008. Primary outcome was SSI, as defined by American College of Surgeons National Surgical Quality Improvement Program. Counseling and global models were built to assess perioperative predictors of superficial incisional SSI and organ/space SSI. Thirty-day cost of SSI was calculated.

RESULTS

Among 1369 EC patients, 136 (9.9%) had SSI. In the counseling model, significant predictors of superficial incisional SSI were obesity, American Society of Anesthesiologists (ASA) score >2, preoperative anemia (hematocrit <36%), and laparotomy. In the global model, significant predictors of superficial incisional SSI were obesity, ASA score >2, smoking, laparotomy, and intraoperative transfusion. Counseling model predictors of organ/space SSI were older age, smoking, preoperative glucose >110 mg/dL, and prior methicillin-resistant Staphylococcus aureus (MRSA) infection. Global predictors of organ/space SSI were older age, smoking, vascular disease, prior MRSA infection, greater estimated blood loss, and lymphadenectomy or bowel resection. SSI resulted in a $5447 median increase in 30-day cost.

CONCLUSIONS

Our findings are useful to individualize preoperative risk counseling. Hyperglycemia and smoking are modifiable, and minimally invasive surgical approaches should be the preferred surgical route because they decrease SSI events. Judicious use of lymphadenectomy may decrease SSI. Thirty-day postoperative costs are considerably increased when SSI occurs.

Incidence and cost of anal, penile, vaginal and vulvar cancer in Denmark

Background

Besides being a causative agent for genital warts and cervical cancer, human papillomavirus (HPV) contributes to 40-85% of cases of anal, penile, vaginal and vulvar cancer and precancerous lesions. HPV types 16 & 18 in particular contribute to 74-93% of these cases. Overall the number of new cases of these four cancers may be relatively high implying notable health care cost to society. The aim of this study was to estimate the incidence and the health care sector costs of anal, penile, vaginal and vulvar cancer.

Methods

New anogenital cancer patients were identified from the Danish National Cancer Register using ICD-10 diagnosis codes. Resource use in the health care sector was estimated for the year prior to diagnosis, and for the first, second and third years after diagnosis. Hospital resource use was defined in terms of registered hospital contacts, using DRG (Diagnosis Related Groups) and DAGS (Danish Outpatient Groups System) charges as cost estimates for inpatient and outpatient contacts, respectively. Health care consumption by cancer patients diagnosed in 2004–2007 was compared with that by an age- and sex-matched cohort without cancer. Hospital costs attributable to four anogenital cancers were estimated using regression analysis.

Results

The annual incidence of anal cancer in Denmark is 1.9 per 100,000 persons. The corresponding incidence rates for penile, vaginal and vulvar cancer are 1.7, 0.9 and 3.6 per 100,000 males/females, respectively. The total number of new cases of these four cancers in Denmark is about 270 per year. In comparison, the total number of new cases cervical cancer is around 390 per year. The total cost of anogenital cancer to the hospital sector was estimated to be 7.6 million Euros per year. Costs associated with anal and vulvar cancer constituted the majority of the costs.

Conclusions

Anogenital cancer incurs considerable costs to the Danish hospital sector. It is expected that the current HPV vaccination program will markedly reduce this burden.

Techniques for estimating health care costs with censored data: an overview for the health services researcher

Objective

The aim of this study was to review statistical techniques for estimating the mean population cost using health care cost data that, because of the inability to achieve complete follow-up until death, are right censored. The target audience is health service researchers without an advanced statistical background.

Methods

Data were sourced from longitudinal heart failure costs from Ontario, Canada, and administrative databases were used for estimating costs. The dataset consisted of 43,888 patients, with follow-up periods ranging from 1 to 1538 days (mean 576 days). The study was designed so that mean health care costs over 1080 days of follow-up were calculated using naïve estimators such as full-sample and uncensored case estimators. Reweighted estimators – specifically, the inverse probability weighted estimator – were calculated, as was phase-based costing. Costs were adjusted to 2008 Canadian dollars using the Bank of Canada consumer price index (http://www.bankofcanada.ca/en/cpi.html).

Results

Over the restricted follow-up of 1080 days, 32% of patients were censored. The full-sample estimator was found to underestimate mean cost ($30,420) compared with the reweighted estimators ($36,490). The phase-based costing estimate of $37,237 was similar to that of the simple reweighted estimator.

Conclusion

The authors recommend against the use of full-sample or uncensored case estimators when censored data are present. In the presence of heavy censoring, phase-based costing is an attractive alternative approach.

Direct medical costs of constipation from childhood to early adulthood: a population-based birth cohort study

BACKGROUND

Although direct medical costs for constipation-related medical visits are thought to be high, to date there have been no studies examining whether longitudinal resource use is persistently elevated in children with constipation. Our aim was to estimate the incremental direct medical costs and types of health care use associated with constipation from childhood to early adulthood.

METHODS

A nested case-control study was conducted to evaluate the incremental costs associated with constipation. The original sample consisted of 5718 children in a population-based birth cohort who were born during 1976 to 1982 in Rochester, MN. The cases included individuals who presented to medical facilities with constipation. The controls were matched and randomly selected among all noncases in the sample. Direct medical costs for cases and controls were collected from the time subjects were between 5 and 18 years of age or until the subject emigrated from the community.

RESULTS

We identified 250 cases with a diagnosis of constipation in the birth cohort. Although the mean inpatient costs for cases were $9994 (95% Confidence interval [CI] 2538-37,201) compared with $2391 (95% CI 923-7452) for controls (P = 0.22) during the time period, the mean outpatient costs for cases were $13,927 (95% CI 11,325-16,525) compared with $3448 (95% CI 3771-4621) for controls (P < 0.001) during the same time period. The mean annual number of emergency department visits for cases was 0.66 (95% CI 0.62-0.70) compared with 0.34 (95% CI 0.32-0.35) for controls (P < 0.0001).

CONCLUSIONS

Individuals with constipation have higher medical care use. Outpatient costs and emergency department use were significantly greater for individuals with constipation from childhood to early adulthood.

Longitudinal direct medical costs associated with constipation in women

BACKGROUND

Although direct medical costs for constipation-related medical visits are thought to be high, to date, there have been no studies examining longitudinal resource utilisation in adults with constipation.

AIM

To estimate the incremental direct medical costs associated with constipation in women.

METHODS

This is a nested case-control study. The study population consisted of all mothers of 5718 children in the population-based birth cohort born during 1976-1982 in a community. The cases presented to the medical facilities with constipation. The controls were randomly selected and matched to cases in a 2:1 ratio. Direct medical costs for constipated women and controls were collected for the years 1987-2002.

RESULTS

We identified 168 women with a diagnosis of constipation. The total direct medical costs over the 15-year period for constipated subjects were more than double those of controls [$63 591 (95% CI: 49 786-81 396) vs. $24 529 (95% CI: 20 667-29 260)]. The overall out-patient costs for constipated women were $38 897 (95% CI: 31 381-48 253) compared to $15 110 (95% CI: 12 904-17 781) for controls. The median of annual out-patient visits for constipated women was 0.16 compared to 0.11 for controls.

CONCLUSION

Women with constipation have significantly higher medical care utilisation and expenditures compared with women without constipation.

Long-term direct costs before and after proctocolectomy for ulcerative colitis: a population-based study in Olmsted County, Minnesota

PURPOSE

This study was designed to test the hypothesis that patients undergoing definitive surgery for chronic ulcerative colitis have reduced direct medical costs after, as compared with before, total proctocolectomy.

METHODS

A population-based cohort of patients who underwent proctocolectomy for ulcerative colitis from 1988 to 2007 was identified using the Rochester Epidemiology Project. Total direct healthcare costs were estimated from an administrative database. The primary outcome was the observed cost difference between the two-year period before surgery and the two-year period after a surgery/recovery period (surgery + 180 days). Statistical significance was assessed using paired t-tests and bootstrapping methods. Demographic data were presented as median (interquartile range) or frequency (proportion). Mean costs are reported in 2007 constant dollars.

RESULTS

Sixty patients were Olmsted County, Minnesota, residents at the time of surgery and for the entire period of observation. Overall 40 patients (66%) were men, median age was 42 (range, 31-52) years, and duration of median colitis was four (range, 1-11) years. Operations included ileal pouch-anal anastomosis (n = 45, mean cost of surgery/recovery period = $50,530) and total proctocolectomy with Brooke ileostomy (n = 15, mean cost of surgery/recovery period = $39,309). In the pouch subgroup, direct medical costs on average were reduced by $9,296 (P < 0.001, bootstrapped 95% confidence interval: $324-$15,628) during the two years after recovery. In the Brooke ileostomy subgroup, direct medical costs on average were reduced by $12,529 (P < 0.001, bootstrapped 95% confidence interval: $6,467-$18,688) in the two years after recovery.

CONCLUSION

Surgery for chronic ulcerative colitis resulted in reduced direct costs in the two years after surgical recovery. These observations suggest that surgical intervention for ulcerative colitis is associated with long-term economic benefit.

Impact of comorbid depression or anxiety on patterns of treatment and economic outcomes among patients with diabetic peripheral neuropathic pain

OBJECTIVE

The objective of this retrospective analysis was to assess the correlation of comorbid depression and/or anxiety to patterns of treatment, healthcare utilization, and associated costs among diabetic peripheral neuropathic pain (DPNP) patients, employing a large US administrative claims database.

RESEARCH DESIGN AND METHODS

Patients under age 65 with commercial insurance and patients aged 65 and older with employer-sponsored Medicare supplemental insurance were selected for the study if they had at least one diagnosis of DPNP in 2005. The first observed DPNP claim was considered the 'index date.' All individuals had a 12-month pre-index and 12-month follow-up period. For both populations, two subgroups were constructed for individuals with depression and/or anxiety (DPNP-DA cohort) or without these disorders (DPNP-only cohort). Patients' demographic characteristics, clinical characteristics, and medication use were compared over the pre-index period. Healthcare expenditures and resource utilization were measured for the post-index period. Two-part models were used to examine the impact of comorbid depression and/or anxiety on healthcare utilization and costs, controlling for demographic and clinical characteristics.

RESULTS

The study identified 11,854 DPNP-only and 1512 DPNP-DA patients in the Medicare supplemental cohort, and 11,685 and 2728 in the commercially insured cohort. Compared to DPNP-only patients over the follow-up period, a significantly higher percentage of DPNP-DA patients were dispensed pain and DPNP-related medication. All components of healthcare utilization, except home healthcare visits and physician office visits, were more likely to be provided to DPNP-DA patients versus the DPNP-only cohort (all p < 0.01). Controlling for differences in demographic and clinical characteristics, DPNP-DA patients had significantly higher total costs than those of DPNP-only patients for Medicare ($9134, p < 0.01) and commercially insured patients ($11,085, p < 0.01).

LIMITATIONS

Due to the use of a retrospective administrative claims database, limitations of this study include the potential for selection bias between study cohorts, mis-identification of DPNP and/or depression, and inability to assess indirect costs as well as use and cost of over-the-counter medications.

CONCLUSIONS

These findings indicate that the healthcare costs were significantly higher for DPNP patients comorbid with depression and/or anxiety relative to those without such disorders.

Healthcare use and costs in patients with chronic bronchitis initiating maintenance therapy with fluticasone/salmeterol vs other inhaled maintenance therapies

OBJECTIVE

To compare risk of hospitalization or emergency department (ED) visit and healthcare costs in patients with chronic bronchitis initiating inhaled maintenance therapy with fluticasone propionate/salmeterol 250/50 mcg combination (FSC) versus other inhaled maintenance therapies.

DESIGN AND METHODS

This retrospective cohort study assessed 9,217 patients from the PharMetrics administrative claims database enrolled from July 1997 to January 2005. Study subjects were persons with medical claims with diagnoses of chronic bronchitis (ICD-9-CM 491.xx) who also had pharmacy claims for FSC, salmeterol (SAL), inhaled corticosteroid (ICS), ipratropium (IPR), or ipratropium/albuterol combination (IAC). Persons with <12 months of continuous eligibility after the first prescription for initial maintenance therapy ("index date") were excluded as were those receiving fluticasone propionate/salmeterol 100/50 mcg or 500/50 mcg (not indicated for patients with chronic bronchitis). For remaining persons, time to first hospitalization or ED visit during follow-up was compared for those receiving FSC versus other therapies using Cox proportional hazards regression. Healthcare costs during the first 12 months of follow-up were analyzed using generalized linear model regression.

RESULTS

Receipt of FSC as initial inhaled maintenance therapy for chronic bronchitis (n = 1361) was associated with 41% lower risk of COPD-related hospitalization or ED visit compared with IPR (n < 1316) (p < 0.001). Adjusted costs of COPD-related hospitalization/ED visit were $507 (95% CI $218-$1083) less with FSC than IPR. However, patients receiving FSC had $261 (95% CI $205-$322) higher COPD-related pharmacy costs than those receiving IPR. Total COPD-related costs were $90 lower with FSC than IPR although this difference was not significant (95% CI $330-$443). Compliance, as measured by medication possession ratio, was 12% greater with FSC compared with IPR (p < 0.05). Comparisons of FSC with IAC yielded generally similar results. The limitations of the study are similar to those of other observational studies of secondary data regarding potential misclassification and omitted variable bias and residual confounding.

CONCLUSIONS

In persons with chronic bronchitis, initial maintenance therapy with FSC 250/50 mcg was associated with improved outcomes versus ipratropium-based therapy and although FSC was associated with greater pharmacy costs, it did not significantly increase total costs of COPD-related care.

Costs of stroke using patient-level data: a critical review of the literature

BACKGROUND AND PURPOSE

With decision-analytic models becoming more popular to assess the cost-effectiveness of health care interventions, the need for robust estimates on the costs of cerebrovascular disease is paramount. This study reports the results from a literature review of the costs of cerebrovascular diseases, and assesses the quality of the published evidence against a set of defined criteria.

METHODS

A broad literature search was conducted. Those studies reporting mean/median costs of cerebrovascular diseases derived from patient-level data in a developed country setting were included. Data were abstracted using standardized reporting forms and assessed against 4 predefined criteria: use of adequate methodologies, use of a population-based study, inclusion of premorbid resource use, and reporting of costs by different patient subgroups.

RESULTS

A total of 120 cost studies were identified. The cost estimates of stroke were compared by taking into account the effects of inflation and price differentials between countries. Average costs of stroke ranged from $468 to $146 149. Differences in costs were also found within country, with estimates in the USA varying 20-fold. Although the costing methodologies used were generally appropriate, only 5 studies were based on population-based studies, which are the gold standard study design when comparing incidence, outcome, and costs.

CONCLUSIONS

This review showed large variations in the costs of stroke, mainly attributable to differences in the populations studied, methods, and cost categories included. The wide range of cost estimates could lead to selection bias in secondary health economic analyses, with authors including those costs that are more likely to produce the desired results.

Design of the economic evaluation for the Interventional Management of Stroke (III) trial

RATIONALE

Stroke is a common and costly condition where an effective early treatment may be expected to affect patients' future quality of life, the cost of acute medical treatment, and the cost of rehabilitation and any supportive care needed for their remaining lifetime. To assist in informing discussions on early adoption of potential treatments, economic analyses should accompany investigations that seek to improve outcomes for stroke patients.

AIMS

The primary aim is to assess whether i.v./i.a. rt-PA therapy is cost-effective at 3 months compared with i.v. rt-PA, and provides cost-savings or is cost-neutral by 12 months. Design Cost-effectiveness of the two treatment arms will be measured at months 3, 6, 9, and 12. Cost-effectiveness will be calculated using 1.standard cost-effectiveness methodology (incremental cost-effectiveness ratios), and 2.an econometric model to assess multiple outcome measures while controlling for multiple subject and treatment-related factors that are known to affect both outcomes and costs.

STUDY OUTCOMES

Total cost for the initial hospitalization of treating stroke subjects randomized to either i.v./i.a. or i.v. rt-PA treatment arms will be measured, as will differences in types of resource utilization over 12 months between the two arms of the trial. Quality-of-life data (EuroQol EQ-5D) will be collected over a 12-month period and quality-adjusted life years will be used as a morbidity-adjusted measure of effectiveness. Subgroup analyses will include dichotomized NIH Stroke Scale (<20, >or=20), country, time between onset and randomization, and i.a. devices.

The impact of home hospitalization on healthcare costs of exacerbations in COPD patients

Home-hospitalization (HH) improves clinical outcomes in selected patients with chronic obstructive pulmonary disease (COPD) admitted at the emergency room due to an exacerbation, but its effects on healthcare costs are poorly known. The current analysis examines the impact of HH on direct healthcare costs, compared to conventional hospitalizations (CH). A randomized controlled trial was performed in two tertiary hospitals in Barcelona (Spain). A total of 180 exacerbated COPD patients (HH 103 and CH 77) admitted at the emergency room were studied. In the HH group, a specialized respiratory nurse delivered integrated care at home. The average direct cost per patient was significantly lower for HH than for CH, with a difference of euro 810 (95% CI, euro 418-1,169) in the mean cost per patient. The magnitude of monetary savings attributed to HH increased with the severity of the patients considered eligible for the intervention.

Estimating the costs for a group of geriatric patients using the Coxian phase-type distribution

The length of stay in hospital of geriatric patients may be modelled using the Coxian phase-type distribution. This paper examines previous methods which have been used to model health-care costs and presents a new methodology to estimate the costs for a cohort of patients for their duration of stay in hospital, assuming there are no further admissions. The model, applied to 1392 patients admitted into the geriatric ward of a local hospital in Northern Ireland, between 2002 and 2003, should be beneficial to hospital managers, as future decisions and policy changes could be tested on the model to investigate their influence on costs before the decisions were carried out on a real ward.

Predicting costs over time using Bayesian Markov chain Monte Carlo methods: an application to early inflammatory polyarthritis

This article focuses on the modelling and prediction of costs due to disease accrued over time, to inform the planning of future services and budgets. It is well documented that the modelling of cost data is often problematic due to the distribution of such data; for example, strongly right skewed with a significant percentage of zero-cost observations. An additional problem associated with modelling costs over time is that cost observations measured on the same individual at different time points will usually be correlated. In this study we compare the performance of four different multilevel/hierarchical models (which allow for both the within-subject and between-subject variability) for analysing healthcare costs in a cohort of individuals with early inflammatory polyarthritis (IP) who were followed-up annually over a 5-year time period from 1990/1991. The hierarchical models fitted included linear regression models and two-part models with log-transformed costs, and two-part model with gamma regression and a log link. The cohort was split into a learning sample, to fit the different models, and a test sample to assess the predictive ability of these models. To obtain predicted costs on the original cost scale (rather than the log-cost scale) two different retransformation factors were applied. All analyses were carried out using Bayesian Markov chain Monte Carlo (MCMC) simulation methods.

Income-related inequality in utilization of health services in Denmark: evidence from Funen County

AIMS

To examine income-related inequity in utilization of healthcare services in Denmark.

METHODS

A health survey of 2,915 respondents in Funen County interviewed in 2000 and 2001 on health status and socioeconomic and sociodemographic characteristics was merged with various computerized registers including inpatient stays, ambulatory visits, contacts in the primary healthcare sector, and prescription medicine. The index of horizontal inequity was used to estimate the degree of horizontal inequity in utilization of healthcare services across income groups, using the indirect method of standardization to control for age, gender, and self-assessed health as a proxy for need. The standardization method rests on the assumption of equal response behaviour across income groups.

RESULTS

The least advantaged with respect to income consume a bigger share of the health services than the most advantaged with the exception of dental treatments where the opposite is true. After standardization for age, gender, and health status there is no significant inequity in use of all healthcare services. However, when it comes to specific healthcare services the least advantaged have a significantly lower share of the medicine consumption and dental treatments than expected.

CONCLUSION

The index of horizontal inequity suggests that the Danish healthcare system is in general equitable. In sectors with a high degree of co-payment some horizontal inequity disfavouring the lower income groups appears to be present.

Income related inequality in prescription drugs in Denmark

PURPOSE

To examine income-related inequity in utilisation of prescription drugs in Funen County, Denmark after a new reimbursement system was implemented.

METHODS

An individual level prescription database was merged with a health survey of 2927 respondents interviewed in 2000 and 2001 about their health status and socio-economic and socio-demographic characteristics. An index of horizontal inequity was used to estimate the degree of inequity in drug utilisation across income groups, using the indirect method of standardisation to control for age, gender and health status as a proxy for need. The results were compared to estimates from a traditional regression analysis.

RESULTS

The least advantaged with respect to income consume a bigger share of the prescription drugs than the most advantaged. After standardisation for age, gender and health status the least advantaged have a lower share of the drug consumption than expected. However, traditional regression analysis showed no signs of an income effect on the level of consumption of prescription drugs.

CONCLUSIONS

The index of horizontal inequity suggests that some horizontal inequity favouring the better off is present. However, the results deviate from what can be found by traditional regression analysis.

Surgeons' economic profiles: can we get the "right" answers?

Hospitals and payers use economic profiling to evaluate physician and surgeon performance. However, there is significant variation in the data sources and analytic methods that are used. We used information from a hospital's cardiac surgery and cost accounting information systems to create surgeon economic profiles. Three scenarios were examined: (1) surgeon modeled as fixed effect with no patient-mix adjustment; (2) surgeon modeled as fixed effect with patient-mix adjustment; (3) and surgeon modeled as random effect with patient-mix adjustment. We included 574 patients undergoing coronary artery bypass surgery at Baptist Medical Center, Oklahoma City, OK between July 1, 1995 and April 30, 1996. We found that profiles reporting unadjusted average surgeon costs may incorrectly identify high- and low-cost outliers. Adjusting for patient-mix differences and treating surgeons as random effects was the preferred approach. These results demonstrate the need for hospitals to reexamine their economic profiling methods.

Smooth quantile ratio estimation with regression: estimating medical expenditures for smoking-attributable diseases

The methodological development of this paper is motivated by a common problem in econometrics where we are interested in estimating the difference in the average expenditures between two populations, say with and without a disease, as a function of the covariates. For example, let Y(1) and Y(2) be two non-negative random variables denoting the health expenditures for cases and controls. Smooth Quantile Ratio Estimation (SQUARE) is a novel approach for estimating Delta=E[Y(1)] - E[Y(2)] by smoothing across percentiles the log-transformed ratio of the two quantile functions. Dominici et al. (2005) have shown that SQUARE defines a large class of estimators of Delta, is more efficient than common parametric and nonparametric estimators of Delta, and is consistent and asymptotically normal. However, in applications it is often desirable to estimate Delta(x)=E[Y(1)|x]--E[Y(2)|x], that is, the difference in means as a function of x. In this paper we extend SQUARE to a regression model and we introduce a two-part regression SQUARE for estimating Delta(x) as a function of x. We use the first part of the model to estimate the probability of incurring any costs and the second part of the model to estimate the mean difference in health expenditures, given that a nonzero cost is observed. In the second part of the model, we apply the basic definition of SQUARE for positive costs to compare expenditures for the cases and controls having 'similar' covariate profiles. We determine strata of cases and control with 'similar' covariate profiles by the use of propensity score matching. We then apply two-part regression SQUARE to the 1987 National Medicare Expenditure Survey to estimate the difference Delta(x) between persons suffering from smoking-attributable diseases and persons without these diseases as a function of the propensity of getting the disease. Using a simulation study, we compare frequentist properties of two-part regression SQUARE with maximum likelihood estimators for the log-transformed expenditures.

Economic endpoints in clinical trials

Healthcare decision makers are increasingly requesting information on the cost and cost-effectiveness of new medicines at the time of product launch. In order to provide this information, data on healthcare resource utilization and, in some cases, costs, may be collected in clinical trials. In this paper, we discuss some of the issues statisticians need to address when it is appropriate to include these economic endpoints in the trial. Several design issues are discussed, including the alternative types of and methods for collecting economic endpoint data, sample size and generalizability. Alternative approaches in the analysis of resource utilization, cost and cost-effectiveness are also presented. Finally, several of the analytic approaches are applied to actual data from a clinical trial.

A comparison of several regression models for analysing cost of CABG surgery

Investigators in clinical research are often interested in determining the association between patient characteristics and cost of medical or surgical treatment. However, there is no uniformly agreed upon regression model with which to analyse cost data. The objective of the current study was to compare the performance of linear regression, linear regression with log-transformed cost, generalized linear models with Poisson, negative binomial and gamma distributions, median regression, and proportional hazards models for analysing costs in a cohort of patients undergoing CABG surgery. The study was performed on data comprising 1959 patients who underwent CABG surgery in Calgary, Alberta, between June 1994 and March 1998. Ten of 21 patient characteristics were significantly associated with cost of surgery in all seven models. Eight variables were not significantly associated with cost of surgery in all seven models. Using mean squared prediction error as a loss function, proportional hazards regression and the three generalized linear models were best able to predict cost in independent validation data. Using mean absolute error, linear regression with log-transformed cost, proportional hazards regression, and median regression to predict median cost, were best able to predict cost in independent validation data. Since the models demonstrated good consistency in identifying factors associated with increased cost of CABG surgery, any of the seven models can be used for identifying factors associated with increased cost of surgery. However, the magnitude of, and the interpretation of, the coefficients vary across models. Researchers are encouraged to consider a variety of candidate models, including those better known in the econometrics literature, rather than begin data analysis with one regression model selected a priori. The final choice of regression model should be made after a careful assessment of how best to assess predictive ability and should be tailored to the particular data in question.

Use of Bayesian Markov Chain Monte Carlo methods to model cost-of-illness data

It is well known that the modeling of cost data is often problematic due to the distribution of such data. Commonly observed problems include 1) a strongly right-skewed data distribution and 2) a significant percentage of zero-cost observations. This article demonstrates how a hurdle model can be implemented from a Bayesian perspective by means of Markov Chain Monte Carlo simulation methods using the freely available software WinBUGS. Assessment of model fit is addressed through the implementation of two cross-validation methods. The relative merits of this Bayesian approach compared to the classical equivalent are discussed in detail. To illustrate the methods described, patient-specific non-health-care resource-use data from a prospective longitudinal study and the Norfolk Arthritis Register (NOAR) are utilized for 218 individuals with early inflammatory polyarthritis (IP). The NOAR database also includes information on various patient-level covariates.

Is rheumatoid arthritis care more costly when provided by rheumatologists compared with generalists?

OBJECTIVE

Controversy surrounds the cost-effectiveness of rheumatologist care compared with generalist care for patients with rheumatoid arthritis (RA). Rheumatologists can provide 2 distinct types of care for RA patients: primary care and specialist care. We sought to examine the relationship between cost and type of care in a population-based cohort of patients with RA.

METHODS

Data regarding specialty of care and use of health services (i.e., total direct medical costs, surgeries, radiographs, laboratory tests, hospital days) were collected from a community sample of 249 patients with RA (defined using the 1987 American College of Rheumatology diagnostic criteria) among Rochester, Minnesota residents > or =35 years of age. In a randomly selected subset of 99 of these RA patients, detailed information on all physician encounters was collected and categorized according to whether or not the care received constituted "primary care" according to the Institute of Medicine definition. Using these data, we evaluated the influence of type of care as well as specialty of provider on utilization. For these analyses, total direct costs included all inpatient and outpatient health care costs incurred by all local providers (excluding outpatient prescription drugs).

RESULTS

The 249 patients with RA (mean age 64 years, 75% women) were followed up for a median of 5.4 years, while the subset of 99 RA patients (mean age 64 years, 77% women) were followed up for a median of 4.7 years. The overall median direct medical costs per person per year were $2,749 and $2,929 for the total cohort and for the subset of 99 patients, respectively. Generalized linear regression analyses (considering all visits of the 249 RA patients) revealed that after adjusting for demographics and disease characteristics, rheumatologist care (compared with nonrheumatologist care) was not associated with higher total direct medical costs (P = 0.85) or more hospital days (P = 0.35), but was associated with slightly more radiographs (P = 0.037) and significantly more laboratory tests (P < 0.0001). When considering only primary care, such care by rheumatologists was, again, not associated with higher total direct medical costs (P = 0.11) or more hospital days (P = 0.69) or more laboratory tests (P = 0.54), but was associated with slightly more radiographs (P = 0.035).

CONCLUSION

Rheumatologist care is not more costly than generalist care for patients with RA. Important differences (especially in the use of laboratory tests) become apparent when the type of care provided as well as the specialty of the provider are considered in the analyses.

An economic cost analysis of oral ganciclovir prophylaxis for the prevention of CMV disease

PURPOSE

The study conducted an economic cost analysis of oral ganciclovir prophylaxis in preventing cytomegalovirus (CMV) disease for AIDS patients in a randomized clinical trial setting.

METHODS

Data were generated from patient interviews, medical records, and case reports from a multi-center, randomized, double-blind, and placebo-controlled pharmacoeconomic study appended to a clinical trial. The outcomes were measured in monthly cost per patient. Various cost functions were tested in the context of sample-selection model (SSM) and two-part model (TPM), and were estimated using both the ordinary least squares (OLS) and the bounded influence estimation (BIE) methods.

RESULTS

The use of informal caregiver services did not differ significantly between patients in the treatment group and those in the placebo group. The OLS estimates for the ganciclovir prophylaxis arm showed a reduced, but statistically insignificant use of formal care in both outpatient and inpatient settings. The BIE results for the ganciclovir prophylaxis arm, in contrast, showed a significant reduction of 27% in hospital cost among hospital users, and 44% among the total sample of AIDS patients. The monthly total cost function also identified a decreasing but insignificant trend due to the treatment effect.

CONCLUSIONS

At the methodological level, this study demonstrated the value of employing more rigorous econometric techniques in identifying subtle treatment effects on cost outcomes from clinical trial data in the economic assessment of medical technologies. At the empirical level, the study concluded that beyond its demonstrated efficacy of preventing CVM disease among AIDS patients, ganciclovir prophylaxis did not lead to additional health care costs, other than the cost of the drug therapy.

Cost function estimation: the choice of a model to apply to dementia

Statistical analysis of cost data is often difficult because of highly skewed data resulting from a few patients who incur high costs relative to the majority of patients. When the objective is to predict the cost for an individual patient, the literature suggests that one should choose a regression model based on the quality of its predictions. In exploring the econometric issues, the objective of this study was to estimate a cost function in order to estimate the annual health care cost of dementia. Using different models, health care costs were regressed on the degree of dementia, sex, age, marital status and presence of any co-morbidity other than dementia. Models with a log-transformed dependent variable, where predicted health care costs were re-transformed to the unlogged original scale by multiplying the exponential of the expected response on the log-scale with the average of the exponentiated residuals, were part of the considered models. The root mean square error (RMSE), the mean absolute error (MAE) and the Theil U-statistic criteria were used to assess which model best predicted the health care cost. Large values on each criterion indicate that the model performs poorly. Based on these criteria, a two-part model was chosen. In this model, the probability of incurring any costs was estimated using a logistic regression, while the level of the costs was estimated in the second part of the model. The choice of model had a substantial impact on the predicted health care costs, e.g. for a mildly demented patient, the estimated annual health care costs varied from DKK 71 273 to DKK 90 940 (US$ 1 = DKK 7) depending on which model was chosen. For the two-part model, the estimated health care costs ranged from DKK 44714, for a very mildly demented patient, to DKK 197 840, for a severely demented patient.

Centre-specific or average unit costs in multi-centre studies? Some theory and simulation

Costing issues are increasingly being addressed in multi-centre studies. In this paper, two methods for collecting costing information are compared within a simulated clinical trial setting. One method estimates average treatment costs by applying unit costs averaged across treatment centres to centre-specific volumes of resource use. The second uses centre-specific information for both the unit costs and the resource volumes, and then averages across centres. Using a pre-specified production relation between the different volumes of resource use, and simulating changes in unit costs, it is shown that these two methods result in statistically different estimates of average treatment costs. This finding holds, regardless of the degree of substitutability between the resource volumes, except when considerable uncertainty surrounds treatment centre responses to relative changes in unit costs. The findings suggest that a more cautious approach should be adopted in the collection, calculation and interpretation of treatment costs in multi-centre studies.

Economic burden of rheumatoid arthritis: a systematic review

OBJECTIVE

To summarize the state of knowledge with regard to the economic impact of rheumatoid arthritis (RA) and to highlight any weaknesses in the work conducted to date, so as to inform future RA cost-of-illness studies.

METHODS

Four computerized literature databases were searched to identify all the literature relevant to this review. Seven elements indicating a quality cost-of-illness study were established and used to appraise the literature identified critically. Where possible, costs reported by the different studies were converted to 1996 US dollars using the consumer price index for medical care.

RESULTS

Total average medical costs were reported to range from US$5720 (UK pound3575) to US$5822 (UK pound3638). Medication constituted between 8 and 24% of total medical costs, physician visits between 8 and 21%, and in-patient stays between 17 and 88%. The average number of days absent from work due to a person's RA was reported to range from 2.7 to 30 days/year.

CONCLUSION

The economic impact of RA in terms of cost was reported to be substantial by all studies reviewed. However, methodological problems meant that discrepancies in the average (per person) annual cost of RA existed across studies.

The use of patient mix-adjusted control charts to compare in-hospital costs of care

We introduce a technique for patient mix-adjusting x charts and compared differences between unadjusted and patient mix-adjusted results. Our data came from coronary artery bypass graft (CABG) surgery patients at Baptist Medical Center, Oklahoma City, Oklahoma. We first developed an unadjusted x control chart to compare monthly changes in CABG surgery costs and then used a published model to patient mix-adjust our x control chart information. Before adjustment, the average log costs for three of ten months were outside the 90% control limit lines, and there was a trend toward increasing costs. After adjustment, two months had average costs outside the 90% lower control limit lines, and the trend toward increasing costs had been explained by differences in patient acuity.