SEARCH

SEARCH BY CITATION

Keywords:

  • Indirect costs;
  • Patient questionnaire;
  • Insurance claims data;
  • Rheumatoid arthritis;
  • Productivity losses

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Objective

To render information on the accuracy of patient-reported indirect cost data compared with payer-derived data of the real indirect costs on a patient-by-patient basis concerning disease-related productivity losses in rheumatoid arthritis (RA).

Methods

The assessment of indirect cost data was part of a clinical, multicenter, randomized RA trial. A total of 234 patients of working age with a diagnosis of RA (according to 1987 American College of Rheumatology criteria) were recruited. Demographics of the cohort were mean age 53 years, mean disease duration 8 years, 76% were women, and all had membership in the regional statutory health insurance plan. Every 3 months corresponding indirect cost data were derived for the cohort from a health economic questionnaire for cost assessment in patients with RA and the payer's database over a period of 18 months. Comparative statistical analyses were performed between patient-reported and insurance claims data.

Results

The mean annual productivity losses due to sick leave amounted to 14 and 17 days per patient (questionnaire versus payer data), and productivity losses due to work disability amounted to 3 days (both); monetary valuation renders overall costs of €1,240 and €1,590, respectively. The difference of 17% in overall productivity losses is not significant. Comparison of productivity losses reveals a strong correlation of r = 0.83 in those due to sick leave and of κ = 0.84 in those due to work disability between questionnaire and payer data.

Conclusion

The comparison of questionnaire and payer data shows that RA patients report their productivity losses adequately. Indirect cost assessment should therefore be included in further RA trials and observational studies.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Current economic evaluations in the field of rheumatic diseases underline the need to include economic data in outcome assessment of clinical trials on rheumatoid arthritis (RA) and other musculoskeletal disorders (1–3). Since many evolving new therapies, such as biologic therapies, are costly, it is important to capture not only clinical data but additionally data on costs and their changes under therapy in order to render information on health economic effectiveness (1).

Reduction or prevention of RA consequences due to therapeutic interventions might lead to indirect cost savings by decreasing productivity losses related to work disability and sick leave. Since these indirect cost components make up >50% of the overall cost of illness in patients with RA (3), their assessment is of special importance.

Studies on original costing data are scarce; hence, most of the current economic analyses are based on mathematic modeling. Valid and easily employable assessment methods are needed to obtain comprehensive direct and indirect cost data covering at least the societal and the patient's perspective in different study settings (e.g., piggy back design).

Cost data assessment has mostly been performed by patient questionnaires so far (4). Different frequently used economic questionnaires in rheumatic conditions have been evaluated in terms of psychometric characteristics (4), showing a wide range of lengths, recall periods, cost units, and the cost domains covered. Only 4 of these (5–8) address the validity of the patient-derived data. Since all studies aim at investigating specific aspects of direct costs, there still is no information on validity and reliability of comprehensive RA cost questionnaires and on indirect cost components in particular.

A recent publication has addressed the validity of patient-derived costing data concerning health care utilization in patients with RA by cross checking information by the health economic patient questionnaire (HEQ-RA) with the database of the health care insurance covering the health care expenses (i.e., direct costs of these patients) (9). Costs for medication and inpatient care accounting for 68% of overall direct costs are reported adequately by the patients. Other cost components showed moderate association of patient- and payer-derived data.

The objective of the present economic investigation is to render information on the validity of patient-reported indirect cost data, as assessed by the HEQ-RA, compared with patient- and payer-derived data on a patient-by-patient basis concerning disease-related productivity losses in RA over a period of 18 months.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Study design and subjects.

The assessment of indirect cost data is part of a multicenter, randomized, controlled, prospective study evaluating the effectiveness of clinical quality management in patients with RA (10). Patients were recruited by 14 outpatient rheumatologists throughout the region of Lower Saxony, Germany. For the present analysis, the following inclusion criteria were applied: diagnosis of RA according to the 1987 American College of Rheumatology (formerly the American Rheumatism Association) classification (11), membership in a major statutory health insurance plan, the “Allgemeine Ortskrankenkasse Niedersachsen,” and the patient had to be of working age (18–65 years in Germany). Indirect cost data are available for 234 patients. Their characteristics and clinical baseline data are shown in Table 1.

Table 1. Patient characteristics and clinical baseline data for 234 subjects of working age with rheumatoid arthritis at study outset*
  • *

    Unless otherwise indicated, values are the number (percentage). ESR = erythrocyte sedimentation rate.

  • Of the 89 currently employed patients.

Demographic variables 
 Female patients177 (76)
 Age, mean ± SE, years53 ± 0.6
Clinical variables 
 Disease duration, mean ± SE, years8 ± 0.5
 Swollen joints, mean ± SE5 ± 0.4
 ESR, mean ± SE, mm/hour15 ± 1.1
 Positive rheumatoid factor148 (63)
 Erosive changes116 (50)
Social variables 
 Currently employed89 (38)
 Early retirement/work disability81 (35)
 Blue collar workers44 (52)
 >9 years education56 (24)

The German social security system.

The German social security system is based on funding by the compulsory health insurance companies (health insurer), the social security pension insurance funds, and the state welfare. During the first 6 weeks of sick leave, an unchanged basic salary is paid by the employer. The start date, duration, and underlying cause of sick leave periods have to be reported to the health care payers by the employers. After week 6 of a sick leave period due to the same disease within 3 years, the health care payers provide about 70–90% of the last income until week 78 of sick leave. Before this 78-week limit is reached, the patients should be referred to a comprehensive, multidisciplinary rehabilitation program if restoration of work ability is feasible. After a maximum of 78 weeks of sick leave, the income depends on the judgement of work disability and on formal criteria of the social security pension insurance fund or the labor office. Patients with positive criteria of work disability receive a disability pension provided by the social security pension insurance fund, and those with positive criteria for unemployment compensation receive payments by the state welfare. Patients not meeting any formal criteria such as a minimum duration of work and payment to the compulsory insurance receive a state welfare at the bare subsistence level.

On the third day of sick leave, the employee has to present a certification by a physician delivering the start date, the diagnosis, and the presumed end of the sick leave period. Thus, a portion of short sick times are not reported to the health insurers.

Data assessment.

Every 3 months a health economic questionnaire for cost assessment in patients with RA (9) has been administered, comprising direct and indirect disease-related cost components over a period of 18 months from October 2000 until March 2002. Corresponding cost data are derived from the Allgemeine Ortskrankenkasse Niedersachsen (health care payer) covering the same time frame. This institution covers medical care for 2.317 million members in the region of Lower Saxony, which is one of 16 regional states in Germany. The data concerning productivity costs are available on a patient-per-patient basis and are matched with the patient-derived data in a single data base.

Indirect RA-related costs incurred by productivity losses due to sick leave periods and cessation of work within the observation period are taken into account. The patient data assessed by questionnaires are reported as sick leave status (sick leave due to RA present or absent) and, in case of sick leave, as cumulated numbers of absence days due to RA certified by a physician (3). Sick leave periods are considered in all 234 persons at working age because unemployed patients are as much a source of incorrect data concerning sick leave assessment as employed patients and should be taken into account in the investigation of the validity of patient-reported data. Additionally, the data on certified sick leave due to RA reported by the employers to the health insurer are assessed.

Work disability assessed by patient questionnaires is considered in 229 patients (5 patients with missing data) in terms of disability status (work disability due to RA present or absent) at the outset of the study and throughout the study and exact date of the start of disability claims. Matching data are recorded by the social security funds and the health insurer. Work disability has been defined as cessation of work receiving disability payment. The information collected by the health insurer can therefore be used as an external data source to cross check the patient-derived data for productivity losses due to sick leave periods and work disability.

In a former publication, a matrix of cost domains for economic evaluation in RA was developed by identifying economic evaluations in a literature search, listing relevant cost categories, and defining a matrix of distinct cost domains avoiding double counting, summarizing related cost categories under a representative heading, and following a valid and feasible taxonomy (12). According to this matrix, all internationally applied health care utilization questionnaires in rheumatic conditions have been analyzed (level of detail, length, type and wording, psychometric characteristics) (4). For the development of the present HEQ-RA, a pool of potential items was generated and applied as a framework for in-depth interviews with patients with RA to determine the preferred level of aggregation of cost domains. The validity of patient-reported resource utilitzation concerning direct cost components of the HEQ-RA has been described by Ruof et al (9).

Economic evaluation.

Productivity losses are estimated from the societal perspective. Both the human capital approach and the human capital approach applying additionally a friction cost period (13) are applied, taking paid work at working age into account. This is in accordance with German guidelines for socioeconomic evaluation (14). Using a friction period means to count productivity losses only within a limited period of time, owing to the fact that the patient's productivity will be replaced assuming that no economy achieves full employment (unemployment rate in Germany, Lower Saxony 10.0% in 2001) (15, 16). The friction period of 58 days is based on statistics of the regional employment offices representing the mean time period until a vacant job reported by an employer to the employment office is procured for a jobless person (17). The time until this new worker is as productive as the person with RA who stopped working is not included in the friction period. The friction method aims at assessing productivity losses in more detail, avoiding overestimation by shifting the focus from the patient to a societal perspective.

In all patients, the number of days of lost productivity due to work disability is cumulated. For the estimation according to the human capital approach, the overall number of days since onset of work disability is applied; for the estimation according to the friction cost approach, the overall number is truncated at 58 days. The sick leave days are evaluated as cumulated numbers of absence days due to RA certified by a physician. These physical units (productivity losses) are then valuated by assuming that a day of lost productivity costs society as much as the average daily German wage estimated by population data. This approach implies that the marginal productivity equals the complete wage costs of a person (13). This average wage is determined by dividing the gross income of all gainfully employed citizens per year by the total number of labor-force participants (14). Using 2001 German population data (18), costs of approximately €74/day of lost productivity arise. The numbers of cumulated days of lost productivity due to either sick leave or work disability are thus multiplied by €74.

Ethical approval and informed consent.

The study protocol has been approved by the ethical committee of the Medical School of Hannover. Data transfer procedures and data protection measures have been approved by the Social Ministry of Lower Saxony. The patients are informed separately about the clinical trial contents and the cost data collection and have signed 2 separate informed consent forms prior to study inclusion.

Statistical analyses.

Data capture and management is carried out with Microsoft ACCESS software (Microsoft, Redmond, WA). For statistical analysis, the SPSS version 10.0 (SPSS, Chicago, IL) is used. Frequencies and mean ± SEM values are calculated for cost data, including the differences of cost data between patient questionnaire and health insurer (concerning sick leave days and work disability days) and patient characteristics. Additionally, 25%, 50%, and 75% quartiles are given for the physical units of cost data. Since the costing variables sick leave and work disability are not normally distributed (tested by Q-Q-plots), the corresponding cost parameters assessed by patient questionnaire and health insurer database are analyzed by cross tables, and the significance of differences is investigated by the Wilcoxon signed rank test and the McNemar test, respectively. Furthermore, Spearman′s rank sum analyses (metric variables) and the Cohen Kappa-correlation (nominal variables) are performed to evaluate the correlation of patient and health insurer data. P values < 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Productivity losses due to sick leave and work disability.

The mean ± SEM annual sick leave periods amount to 17 ± 2.0 days per patient (25%, 50%, and 75% quartiles: 0, 0, and 10, respectively), and productivity losses due to work disability amount to 3 ± 0.1 days according to the health insurer′s data (25%, 50%, and 75% quartiles: 0, 0, and 0, respectively) (Table 2), corresponding to 26 ± 3 and 5 ± 0.2 days of lost productivity regarding the complete observation period. Indirect cost data captured by the patient questionnaire show a similar result. Sick leave periods account for 14 ± 2 days (25%, 50%, and 75% quartiles: 0, 0, and 8, respectively), and work disability for 3 ± 0.2 days (25%, 50%, and 75% quartiles: 0, 0, and 0, respectively), corresponding to 21 ± 3 and 5 ± 0.4 for the entire period. Comparison of the productivity losses based on patient questionnaire data with the health insurer data eludes only minor differences (Table 2). The annual productivity losses due to sick leave periods mentioned above differ ∼18% (14 versus 17 days), those due to work disability differ ∼13% (2.8 versus 3.2 days), and overall costs differ ∼17%. The differences between patient questionnaire and insurer data are not significant according to Wilcoxon's signed rank test or McNemar's test, respectively.

Table 2. Annual productivity losses (PL) and associated indirect costs due to sick-leave periods and work disability in 234 RA working age patients with rheumatoid arthritis followed over 18 months*
 PL days per patient and yearPayer-patient difference of PL daysCosts per patient and year, Euro
Payer dataPatient dataPayer dataPatient data
Sick leave17 ± 2.014 ± 1.72.2 ± 1.01,260 ± 1451,040 ± 130
Work disability3 ± 0.13 ± 0.20.01 ± 0.004240 ± 20200 ± 18
Overall20 ± 2.117 ± 2.02.2 ± 1.01,500 ± 1601,240 ± 150

Regarding the distribution of the differences, the mean ± SEM annual difference of lost productivity days is 2.2 ± 1.0 (Table 2). The frequency of the differences of sick leave days documented by patient questionnaire and insurer show a deviation of more than ±3 days in 20% of all sick leave cases; 8% of these patients overestimated and 12% underestimated their productivity losses. Half of the patients underestimating sick leave periods forget to mention the event at all. This fact proves to be the major source of systematic deviation among patient questionnaire and health insurer data on the number of days of lost productivity.

A cross table (Table 3) comparing the incidence of sick leave periods throughout the observation period shows 83% agreement. False-negative and false-positive results are equally frequent.

Table 3. Cross table of overall sick leave incidences of 234 working age patients with rheumatoid arthritis within the complete observation period*
 Patient: no sick leavePatient: ≥ 1 sick leave period 
  • *

    Values are the number (percentage).

Payer: no sick leave252 (63)30 (8)282 (71)
Payer: ≥ 1 sick leave period37 (9)81 (20)118 (29)
 289 (72)111 (28)400 (100)

The comparison of work disability status within the observation period (Table 4) shows 8% (18 of 229 patients) with a false-negative result being work disabled according to the health insurer's data, but not according to the questionnaire. Of these patients, 3 are housewives and do not mention their disability status, and 15 are naming retirement pension as the reason for pension claims (and not RA). In 7% (16 of 229 patients) a false-positive result was found: the patients reported work disability according to the questionnaire, and the cross check with the health insurer data showed no disability status. Their wrong judgement might be explained by the fact that they have applied for claims that have not yet been decided on.

Table 4. Cross table of work disability status of 229 working age patients with rheumatoid arthritis (5 missing values) within the complete observation period*
 Patient: no work disabilityPatient: work disabled 
  • *

    Values are the number (percentage).

Payer: no work disability119 (52)18 (8)137 (60)
Payer: work disabled16 (7)76 (33)92 (40)
 135 (59)94 (41)229 (100)

Correlation analysis of productivity losses of patient questionnaire and health insurer data.

The correlation of sick leave assessed by patient questionnaire and the health insurer are displayed in Table 5. Results are given for every period (T1–T6) and for the whole observation period. The frequency of sick leave periods reported by the patients (11–16%) and captured by the health insurer (12–18%) are similar (no significant difference in every period). In all periods, the mean numbers of days of sick leave correlate closely among patient questionnaire and health insurer data (P < 0.01). Considering the whole observation period, there is a close correlation of patient questionnaire data and insurer data in terms of the number of sick leave days as well (r = 0.83).

Table 5. Frequencies of sick-leave (SL) periods and number of sick-leave days within 6 3-month periods (T1–T6) in 234 working age patients with rheumatoid arthritis
 T1T2T3T4T5T6Average T1–T6
  • *

    Results of patient-derived data and health insurance data by Spearman correlation coefficient, P < 0.01.

SL incidence patient data, no./total (%)13/83 (16)11/69 (16)23/156 (15)19/177 (11)25/207 (12)22/197 (11)117/905 (13)
SL incidence insurance data, no./total (%)12/83 (15)9/69 (13)28/156 (18)27/177 (15)25/207 (12)28/197 (14)129/905 (14)
SL days/pt. patient data, mean ± SEM5.1 ± 1.64.8 ± 1.94.2 ± 1.32.4 ± 0.83.5 ± 0.92.7 ± 0.83.6 ± 0.5
SL days/pt insurance data, mean ± SEM6.1 ± 2.03.0 ± 1.46.8 ± 1.53.0 ± 0.93.9 ± 1.03.7 ± 0.94.3 ± 0.5
Correlation of SL days*0.830.80.860.660.850.920.83

The correlation analysis of productivity losses due to work disability (according to the friction cost method) among patient questionnaire and health insurer data (Cohen-Kappa coefficient) also reveals a significant association of patient and insurer data (κ = 0.84).

Productivity costs due to sick leave and work disability.

Monetary valuation of the health insurer′s reports on productivity losses renders mean ± SEM overall costs of €1,500 ± €160, with €1,260 ± €145 related to sick leave costs and €240 ± €20 to work disability costs. Regarding the complete observation period, overall costs of €2,250 ± €240 arise.

The corresponding costs assessed by patient questionnaire amount to €1,040 ± €130 for sick leave and €200 ± €18 for work disability, respectively, amounting to annual overall costs of €1,240 ± €150 and of €1,680 ± €225 for the complete observation period. Annual overall costs differ by €260.

Comparison of the friction cost and human capital methodology.

Annual productivity losses due to work disability are assessed by the friction cost method and the human capital approach. The friction cost method renders 638 versus 754 days of lost productivity in all patients (health insurer versus patient data), and the human capital approach 3,780 versus 4,320 days. Since the number of days of lost productivity due to sick leave periods is not altered by different approaches (see methods), the overall productivity losses amount to 4,663 versus 4,124 and 7,805 versus 7,690 days, respectively. The overall productivity losses per patient increase about 1.7 fold applying the human capital approach, rendering annual overall costs of €2,470 and €2,430, respectively.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Assessment of objective cost data by using registers and databases of health insurers is considered the gold standard of measurement of disease-related health care utilization and productivity losses in terms of sick leave and work disability (9). The major drawback of these data sources is their limited accessibility. Difficulties occur due to varying health care payers and insurances of an RA cohort, maintenance of data security, or matching of data sources (clinical and economic). These facts make the assessment of objective cost data a challenge of RA outcome evaluation. The development of valid tools based on patient-derived data for cost assessment plays an important role to integrate economic outcomes in RA outcome studies. The validity of patient-reported productivity losses in patients with rheumatic conditions has not yet been investigated in other studies.

The present economic investigation reveals that productivity cost estimations according to the patient questionnaire HEQ-RA and the health insurer database render comparable results. Regarding the health insurer data as the gold standard, the HEQ-RA underestimates overall costs by 17%. This deviation is lower than expected because the current literature shows a wide range of results concerning indirect cost estimation of RA (19), leading to the assumption that productivity losses might represent a very foggy component of disease-related costs. Analyzing exactly defined cost components such as productivity losses due to sick leave and work disability with a standardized methodology and a defined economic perspective reveals a strong relation between patient- and insurer-derived data considering both cost components. These results indicate that the diverging indirect costs measured in prior cost-of-illness studies can partly be explained by missing standardization of assessment methodology and assessment tools, but not by inadequate patient reports (19). Because patients seem to report their productivity losses adequately, more effort should be made to develop comparable assessment tools and to provide transparency of the assessment methodology applied in future costing studies.

In the present study, a focus was put on components of productivity losses that can be validated by an objective data source (sick leave and work disability), reducing the perspective to societal relevant indirect costs. The productivity losses experienced by housewives or retirees due to impaired daily activities are important components of individual indirect costs (20), but the inclusion of these aspects was not intended in the present investigation. However, the patients not gainfully employed are included in the present analysis, because the adequacy of their reports concerning the components of productivity costs investigated are relevant to determine the validity of patient-reported cost data.

The real overall productivity losses might be underestimated to a certain extent because the questionnaire focuses on sick leave periods that are certified by a physician. Sick leave periods <3 consecutive days do not have to be certified by a physician. However, according to the objective of the present investigation, the data collection is aimed at assessing clearly defined and comparable cost components to evaluate the adequacy of patient-reported data.

The patient′s consent to obtain data from the health insurer was given prior to the study and prior to the completion of the questionnaires. It is quite unlikely that while answering the questionnaires, the patients were aware that the data collected by the questionnaire were to be compared with the insurer data because the economic data assessment was only a part of the complete clinical outcomes assessment. Still, a reporting bias cannot completely be ruled out.

For the comparison of overall productivity costs due to sick leave and work disability with earlier cost-of-illness studies applying the human capital approach and the societal perspective, a detailed discussion of the samples and the methodology is necessary. In the present study, overall productivity costs amount to ∼€2,500 applying the human capital approach. McIntosh (21) reports ∼€3,500, and Kobelt et al (22) report €5,600 per patient and year. Because both other studies have applied models or assumptions for cost estimation based on average data, the differences might be explained by this methodologic reason. Cost estimation of productivity losses based on patient data has been carried out by Clarke et al (23), Magnusson (24), and Merkesdal et al (25), and renders annual indirect costs of €1,330, €17,860, and €10,590, respectively. The result of the study of Clarke et al (23) is extremely low due to the fact that the sample had a high average age of >60 years. Magnusson (24) employed a higher valuation standard compared with the study by Merkesdal et al (25), leading to higher overall indirect costs, but comparing the number of lost productivity days and the sample characteristics, the results of these studies are comparable. In the present study, the amount of productivity losses is lower. This might be related to the fact that all of the patients received optimal specialist treatment before and throughout the study. However, there might be several other reasons leading to deviations of cost estimations, such as different social security systems or different characteristics of study samples.

According to recent methodologic work on indirect cost assessment (19), the friction cost method is recommended. For a better comparability with existing cost-of-illness studies, the human capital approach has been employed alternatively. The productivity losses counted by each approach are reported separately in order to render physical units that can directly be compared with other indirect cost evaluations.

The comparison of patient and insurer data on overall productivity costs and correlation analyses shows that patients are reporting their productivity losses adequately. Only minor deviations can be detected; these have been investigated separately. Identification of systematic failures of the patients′ estimate of productivity losses might help to further improve the assessment methodology.

Although the productivity losses appear to be reported adequately by means of the HEQ-RA in general, there are some limitations of the questionnaire. It has become obvious that patients tend to underestimate the length and number of sick leave periods due to RA (3). In the present investigation, the major reason for this underestimation is sick leave cases not mentioned by the patients. Underestimation of the length of these periods is of minor relevance. Cross checking the data of the HEQ-RA and the insurer's statements about the changes of work disability status within the observation period reveals that 2 patients are reporting RA-related work disability though being gainfully employed. Because both patients reporting work disability did so within the last 3 months of the observation period, a possible reason for the differing data might be that they recently stated disability claims that have not been decided on.

Further improvement of the assessment of productivity losses due to sick leave might be achieved by shortening the recall period. But because the recall period of the HEQ-RA is very short already (3 months), an increase of effectiveness by more frequent data assessment could only be achieved by patient diaries. Such diaries have been employed in other studies, but accompanying problems of this method have to be taken into account. The transfer of the patients′ answers into physical cost units in a database becomes more complicated.

To improve the data on work disability status, a more detailed question about different reasons for work disability and an additional item to check whether work disability was claimed should be added to the HEQ-RA and should be explained to the respondent. Improvement of these 2 issues would probably lead to closer corresponding results.

In summary, the assessment of indirect cost data by the patient questionnaire HEQ-RA has proven to be a valid method to assess disease-related productivity losses in RA because the patients report their productivity losses adequately. Additional assessment of productivity losses and health care utilization should be considered in future RA outcome studies because valid cost estimation is feasible.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  • 1
    Merkesdal S, Ruof J, Mittendorf T, Mau W, Zeidler H. Health economics research in the area of chronic polyarthritis. Z Rheumatol 2002; 61: 219. In German.
  • 2
    Ruof J, Huelsemann JL, Stucki G. Evaluation of costs in rheumatic diseases: a literature review. Curr Opin Rheumatol 1999; 11: 1049.
  • 3
    Mittendorf T, Merkesdal S, Huelsemann JL, von der Schulenburg JM, Zeidler H, Ruof J. Implementing standardized cost categories within economic evaluations in musculoskeletal diseases. Eur J Health Econ 2003; 4: 439.
  • 4
    Ruof J, Merkesdal S, Huelsemann JL, Schoeffski O, Maetzel A, Mau W, et al. Cost assessment instrument in rheumatology: evaluation of applied instrument characteristics. J Rheumatol 2001; 28: 6625.
  • 5
    Clarke AE, Esdaile JM, Bloch DA, Lacaille D, Danoff DS, Fries JF. A Canadian study of the total medical costs for patients with systemic lupus erythematosus and the predictors of costs. Arthritis Rheum 1993; 36: 154859.
  • 6
    Guzman J, Pelosi P, Bombardier C. Capturing health care utilization after occupational low-back pain: development of an interviewer-administered questionnaire. J Clin Epidemiol 1999; 52: 41927.
  • 7
    Goossens ME, Rutten-van Molken MP, Vlaeyen JW, van der Linden SM. The cost diary: a method to measure direct and indirect costs in cost-effectiveness research. J Clin Epidemiol 2000; 53: 68895.
  • 8
    Ritter PL, Stewart AL, Kaymaz H, Sobel DS, Block DA, Lorig KR. Self-reports of health care utilization compared to provider records. J Clin Epidemiol 2001; 54: 13641.
  • 9
    Ruof J, Huelsemann JL, Mittendorf T, Handelmann S, von der Schulenburg JM, Zeidler H, et al. Patient-reported health care utilization in rheumatoid arthritis: what level of detail is required? Arthritis Rheum 2004; 51: 77481.
  • 10
    Hulsemann JL, Mattussek S, Hennig H, Stucki G. Quality management of treatment of rheumatoid arthritis in a rheumatological setting. Z Arztl Fortbild Qualitatssich 2003; 97: 38390. In German.
  • 11
    Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 31524.
  • 12
    Merkesdal S, Ruof J, Huelsemann JL, Schoeffski O, Maetzel A, Mau W, et al. Development of a matrix of cost domains in economic evaluation of rheumatoid arthritis. J Rheumatol 2001; 28: 65761.
  • 13
    Schoeffski O, Glaser P, Schulenburg JM, Graf VD. Health economic evaluation. New York: Springer-Verlag; 1998.
  • 14
    Hannoveraner Konsensus Gruppe. Deutsche empfehlungen zur gesundheitsökonomischen evaluation: Revidierte Fassung des Hannoveraner Konsens. Gesundh okon Qual manag 1999; 4: A625.
  • 15
    Institute for labour market- and vocational research Germany: Unemployed and unemployment rates (2002) Statistisches Bundesamt, Pressestelle; 2002.
  • 16
    Koopmanschap MA: Complementary analyses in economic evaluation of health care [dissertation]. Rotterdam (The Netherlands): Erasmus University; 1994.
  • 17
    Bundesanstalt fuer Arbeit: Strukturanalyse 1993, Nürnberg, Germany: Statistisches Bundesamt, Pressestelle; 1994.
  • 18
    Statistisches Bundesamt. Statistical Yearbook 2001 for the Federal Republic of Germany. Wiesbaden: Metzler Poeschel; 2001.
  • 19
    Merkesdal S, Ruof J, Mittendorf T, Zeidler H, Mau W. Indirect medical costs in the first three years of rheumatoid arthritis: comparison of current methodological approaches. Expert Rev Pharmacoeconomics Outcomes Res 2002; 2: 8994.
  • 20
    Newhall-Perry K, Law NJ, Ramos B, Sterz M, Wong WK, Bulpitt KJ, et al, and the Western Consortium of Practicing Rheumatologists. Direct and indirect costs associated with the onset of seropositive rheumatoid arthritis. J Rheumatol 2000; 27: 115663.
  • 21
    McIntosh E. The cost of rheumatoid arthritis. Br J Rheumatol 1996; 35: 78190.
  • 22
    Kobelt G, Eberhardt K, Jonsson L, Jonsson B. Economic consequences of the progression of rheumatoid arthritis in Sweden. Arthritis Rheum 1999; 42: 34756.
  • 23
    Clarke AE, Zowall H, Levinton C, Assimakopoulos H, Sibley JT, Haga M, et al. Direct and indirect medical costs incurred by Canadian patients with rheumatoid arthritis: a 12 year study. J Rheumatol 1997; 24: 105160.
  • 24
    Magnusson S. Treatment of rheumatoid arthritis: does it affect society′s cost for the disease? Br J Rheumatol 1996; 35: 7915.
  • 25
    Merkesdal S, Ruof J, Schoffski O, Bernitt K, Zeidler H, Mau W. Indirect medical costs in early rheumatoid arthritis: composition of and changes in indirect costs within the first three years of disease. Arthritis Rheum 2001; 44: 52834.