- Email: [email protected]
Re-interventions after restoring teeth—Mining an insurance database
Accepted Manuscript Title: Re-interventions after restoring teeth—Mining an insurance database Author: Michael Raedel Andrea Hartmann Heinz-Werner Priess Steffen Bohm Stefanie Samietz Ioannis Konstantinidis Michael H. Walter PII: DOI: Reference:
S0300-5712(16)30243-3 http://dx.doi.org/doi:10.1016/j.jdent.2016.11.011 JJOD 2705
To appear in:
Journal of Dentistry
Received date: Revised date: Accepted date:
25-4-2016 23-10-2016 22-11-2016
Please cite this article as: Raedel Michael, Hartmann Andrea, Priess Heinz-Werner, Bohm Steffen, Samietz Stefanie, Konstantinidis Ioannis, Walter Michael H.Reinterventions after restoring teeth—Mining an insurance database.Journal of Dentistry http://dx.doi.org/10.1016/j.jdent.2016.11.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Re-interventions after restoring teeth – mining an insurance database
Short title:
Re-interventions after restoring teeth
Authors: Michael Raedela,
dentist /research associate [email protected]
Andrea Hartmannb,
biomedical statistician
[email protected]
Heinz-Werner Priessb biomedical statistician
[email protected]
Steffen Bohmb,
[email protected]
chief executive officer c
Stefanie Samietz
dentist /research associate [email protected]
Ioannis Konstantinidisa
dentist/research associate
[email protected] Michael H. Waltera,
professor and chair
[email protected]
Affiliations: a
Department of Prosthetic Dentistry
Technische Universität Dresden, Faculty of Medicine Carl Gustav Carus Fetscherstraße 74, 01307 Dresden, Germany b
AGENON
Gesellschaft für Forschung und Entwicklung im Gesundheitswesen Kaunstraße 21, 14163 Berlin, Germany c
Policlinic of Prosthetic Dentistry, Gerodontology and Biomaterials,
Center of Oral Health, University Medicine Greifswald, Rotgerberstr. 8, 17475 Greifswald, Germany
Corresponding author: Michael Raedel Department for Prosthetic Dentistry Faculty of Medicine Carl Gustav Carus Fetscherstraße 74 01307 Dresden Telephone:
+49 351 458-4231; Telefax: +49 351 458-5314
E-mail: [email protected]
Abstract Objectives: The aim of this study was to examine re-interventions after restorative treatment.
Methods: The data was collected from the digital database of a major German national health insurance company. Only permanent teeth were observed. Placing a permanent restoration other than a crown regardless of involved surfaces and material was the study intervention. The data did not allow for a differentiation between fillings and inlays that were estimated only a very small portion of the restorations. Success was defined as not undergoing any restorative re-intervention with fillings or inlays on the same tooth (primary outcome) and assessed with Kaplan-Meier survival analyses over four years. An additional analysis was conducted rating “crowning” and “extraction” of respective teeth as target events. Differences were tested with the Log-Rank-test. A multivariate Cox regression analyses was carried out.
Results: A total of 17,024,344 restorations placed in 4,825,408 anterior teeth and 9,973,177 posterior teeth could be traced. Focussing on the primary outcome re-intervention, the cumulative four-year success rate was 69.9 % for one surface restorations, 74.8 % for two surface restorations, 66.6 % for three surface restorations and 61.0 % for four surface and more extended restorations. These differences were significant (p<0.0001). Focussing on all three target events re-intervention, crowning and extraction, the cumulative four-year success rate was 66.1 % for one surface restorations, 67.5 % for two surface restorations, 63.0 % for three surface restorations and 55.8 % for four surface and more extended restorations. The number of restoration surfaces as well as the tooth position remained significant in the multivariate Cox regression.
Conclusions: The sustainability of restorative dental treatment under the terms and conditions of the German national health insurance system shows room for improvement.
From a public health perspective, special focus should be laid on primary and secondary prevention to minimize the restorative treatment need.
Clinical significance statement: This study shows that re-interventions are observed regularly after restorative treatment. Therefore, preventive and restorative strategies should be revisited and optimised.
key words:
dental restoration (permanent), general practice (dental), health services
research, operative dentistry, public health, treatment outcome
Introduction Caries is still one of the most prevalent diseases in dentistry worldwide. The unrecoverable loss of tooth structure caused by caries causes high costs for dental treatment in a long-term perspective. From the perspective of patients as well as from a public health perspective the sustainability of treatment is an important aspect in dentistry. There are treatment options that generally provide good sustainability because of their high long-term success or survival rates, for example single crowns (1). Regarding dental fillings and inlays, published success or survival rates differ. Especially when focussing on outcomes from general practice, data are rare.
Clinical studies and systematic reviews for different tooth types are available (2-6) mainly indicating favourable outcomes. However, it is quite obvious that results from clinical research are not representative for outcomes from general practice. There are results from single practices (7) and practice networks (8) delivering important data. Especially large practice based research networks (PBRN) using digital techniques for collecting data recently became an important addition to clinical research. Participating in research projects however changes attitudes and treatment decisions (9). Therefore, results from single practices or PBRN come closer to general practice but might still not be representative. At
the moment, the most reliable option for measuring outcomes directly and comprehensively from general practice can be realised in mining databases of public health care systems or public health insurances. These databases being originally created for payment of dentists, their scientific use may be often limited and methodologically restricted. However, these databases often provide treatment courses of several million patients.
Concerning restorative treatment, studies using massive data are quite rare. Highest case numbers were reached in studies conducted by Burke & Lucarotti using data from the British National Health Service (10, 11). In a four-year period, they found success rates roughly between 80 % for single surface amalgam restorations and 60 % for glass ionomer and composite resin restorations. A similar study based on private insurance data from the United States evaluated 300,753 patient cases (11). Success rates ranged between 60 % and 92 % at seven years. Other studies analysing treatment data from public health dental offices comprise much lower case numbers (13-14).
Therefore, the aim of this study was to extend the knowledge about the sustainability of restorative treatment under general practice conditions on the basis of a large data set. This data set consists of 8.6 million members of a German national health insurance company.
Materials and Methods The study was based on routine data of a major German national health insurance company (BARMER GEK, Berlin, Germany). This insurance company publishes annual health care reports (15, 16). In this context, the study group had access to the company’s data warehouse. The study design was approved by the responsible local ethics board (EK 288072015).
Fee codes and treatment dates for every single treatment step were available and allowed for tracing clinical courses on a day-count basis. Data were available for a four-year period from 01.01.2010 until 31.12.2013. Only data sets of patients that had been a member of the
insurance company for the whole four-year observation period entered the analysis. Because of systematically missing data, some specific German regions had to be excluded.
The placement of a permanent dental restoration other than a crown into a permanent tooth was defined as the initial study intervention. This marks the date of study intervention for the respective tooth. The claims data did not allow for a differentiation between fillings and inlays. The insurance system concedes different additional payments depending on the type of restoration that were not accessible for analysis. Inlays, however, were estimated only a very small portion of the restorations. Two independent survival analyses were calculated: In the first analysis, a re-intervention by placing a filling or an inlay on the respective tooth was defined as the primary outcome (primary re-intervention). This re-intervention can be a complete or partial renewal of the original restoration or a restoration independent from the original one. Within the primary outcome analysis, the occurrence of the events “crowning of the respective tooth” and “extraction of the respective tooth” were not rated as a target event but led to censoring. In a second analysis, all re-interventions were rated as target events. These re-interventions comprised the placing of a new filling or inlay in the respective tooth, the crowning of the respective tooth or the extraction of the respective tooth. Because permanent dental restorations may be used for building up teeth to be crowned, crowning was not counted as target event within 60 days after the restoration had been placed. The respective tooth type (anterior teeth, posterior teeth) was obtained from the database. The size of the dental restoration was categorized by the number of involved tooth surfaces (one, two, three, four and more extended). For all teeth that underwent no re-intervention, the observation period ended on December 31st 2013.
Survival analyses were conducted according to the Kaplan-Meier method. Differences between survival functions were first tested with the Log-Rank-Test. Additionally, a multivariate Cox-Regression analysis was carried out. The level of significance was set to p <
0.05. Survival analyses were conducted separately for each outcome on a day count basis. The software SAS (Statistical Analysis System, SAS Institute, Cary, NC, USA) was used for data preparation. The software R (available from http://www.r-project.org) with the add-on package “survival” was used for statistical analyses.
Results The study sample comprised 14,798,585 teeth with 17,024,344 restorations in 3,924,245 patients. The mean age was 51.4 years (Figure 1). A total of 4,825,408 anterior teeth and 9,973,177 posterior teeth were traced after having undergone a study intervention. The study intervention comprised 4,539,634 one surface restorations, 6,451,138 two surface restorations, 3,840,656 three surface restorations and 2,192,916 four surface and more extended restorations.
Focussing on the primary outcome re-intervention, one surface restorations showed a cumulative success rate of 69.9 % at four years. This value corresponds to an annual reintervention rate (ARR) of 7.5 %. Two surface, three surface and four surface and more extended restorations showed cumulative success rates of 74.8 % (ARR = 6.3 %), 66.6 % (ARR = 8.35 %) and 61.0 % (ARR = 9.75 %) at four years (Figure 2). The respective cumulative success rates at two years were 82.8 % (one surface), 85.2 % (two surface), 81.1 % (three surface) and 76.3 % (four surface and more extended). The Log-Rank test showed a significant difference between the survival functions (p<0.0001). For the secondary analysis comprising all three outcomes, the four-year success rates were 66.1 % for one surface restorations, 67.5 % for two surface restorations, 63.0 % for three surface restorations and 55.8 % for four surface and more extended restorations (Figure 3).
For the primary outcome, the mean observation period was 662.2 days for one surface restorations, 698.0 days for two surface restorations, 650.1 days for three surface
restorations and 622.4 days for four and more surface restorations. The respective median values were 638 days, 707 days, 616 days and 568 days. At two years, a number of 1,974,742 one surface restorations, 3,120,992 two surface restorations, 1,612,838 three surface restorations and 861,800 four surface and more extended restorations were still under observation. The respective numbers at the time of the last event were 5,821, 14,123, 9,394 and 10,065. The last events occurred between day 1,449 and day 1,456.
Comparing four-year success rates for equal restoration sizes in anterior and posterior teeth revealed small but significant differences. One surface restorations and four surface and more extended performed worse in anterior teeth compared to posterior teeth (Table 1). Two surface restorations and three surface restorations performed better in anterior teeth compared to posterior teeth.
In the multivariate Cox regression analysis, a higher number of restoration surfaces as well as a posterior tooth were found to be significantly associated with a higher risk for a reintervention (p<0.0001). Detailed results are shown in table 2.
Discussion Focussing on the primary outcome re-intervention, the presented study revealed cumulative success rates between 60 % and 70 % at four years based on an exceptional high number of cases. However, some limitations have to be taken into account in order to facilitate an appropriate discussion and to avoid over-interpretation.
Limitation The major limitation of this study was the data source being not a clinical documentation but fee code based data provided by an insurance company. On the one hand, the database comprised the complete number of all conducted treatments for the 8.6 million members of
this national health insurance company representing 10.5 % of the German population in December 2013. On the other hand, there was no additional clinical information except for the documented treatment fee codes which are digitally submitted from the dentist to the insurance company without involving the patient. However, given appropriate interpretation, the value of the respective results is unquestionable. Health Services Research and the use of massive data resources for evaluating outcomes from general dental practice are gaining increasing importance within the scientific community. In this context, the aspect of data quality has to be mentioned. Fee code based data may be compromised by a mix-up of teeth, surfaces and fee codes although the respective errors may to a certain extent compensate for one another. However, a potential bias has to be taken into account but is all in all judged minor.
Dental restorations were detected based on respective treatment fee codes. These codes did not deliver information about materials (composite resins, amalgam, glass ionomer cements, different inlay materials and other) or techniques. An estimation of the distribution of the material for posterior teeth can be derived from the results of a limited explorative survey conducted by the study group independently from this analysis. Within this survey, patients were asked about the materials used for restorations in their first molars. Composite materials showed a percentage of 44 % followed by amalgam with 20 %. 21 % of the patients did not remember. All other materials were below 10 %. Comparable studies based on data from England and Wales were able to detect differences between materials (10). The lack of respective data is a second major limitation.
Re-interventions were evaluated on the tooth level. This unusual approach was chosen with respect to the sustainability of restorative dental treatment. Of course, the presented results will somehow overestimate re-intervention rates on a restoration level. Furthermore, modern concepts for restoring teeth even prefer small restorations and repairs over larger restorations in order to save tooth structure (17-19). However, from the patients’ and the
health economics perspectives, the sustainability of restorative treatment is an important issue. After placement of a restoration other than a crown, a tooth should be seen as completely restored in a medium-term perspective. For evaluating the sustainability of the treatment concept, this study aimed at detecting each re-intervention on the same tooth within the four year period. This is feasible although the re-intervention might not directly affect the initial restoration. Due to this methodology, the presented results cannot be directly compared with restoration based success or survival rates.
The placement of a new restoration other than a crown at the same tooth was chosen as the primary outcome. It can be seen as the most likely event associated with the previous placement of a dental restoration. Crowning or extracting a tooth might also be associated with the previous placement of a dental restoration but several other reasons come into question. Crowning and extraction therefore served as secondary outcomes. Due to these different indications, a combined analysis was considered inappropriate. The four-year observation period was limited by the availability of the data. It can be judged acceptable to evaluate the sustainability in a medium term perspective. Regarding the statistical methods, the Kaplan-Meier survival analysis can be seen as a common method for describing outcomes after interventions in medical and dental studies. It is an accepted method for outcome analyses on the basis of massive data (10, 11, 20-24).
Interpretation Interpreting the presented results is difficult due to the lack of comparable studies. In this dimension, there is no study available focussing on the outcome of anterior and posterior teeth on a tooth level. The vast majority of studies or reviews only focus on posterior teeth (37, 10, 11, 13, 14, 25-28). Nevertheless, we see two major aspects for interpretation. One aspect is the range of the presented success rates itself in comparison to success/survival rates from other studies. The other aspect can be seen in the significant differences this study revealed between success/survival rates for different tooth types and restoration sizes.
Focussing on the primary outcome re-intervention, the overall success rates at four years were roughly found between 60 % and 75 %. That is, at least every 4th tooth underwent a reintervention within four years after a permanent restoration. Results from practice based research networks tend to be superior (8). Furthermore, taking all outcomes into account, the results raise even more questions. But the presented results are not representing restoration based success or survival rates. The performance of a single restoration might be underestimated by this approach. Therefore, the overall judgement is difficult and somehow controversial.
There are several reasons for considering our results to be acceptable. The major argument might be seen in the contemporary preference of restoring defects as small as possible (1719). This might tend to a higher number of interventions itself. A second point is that socioeconomic factors might affect treatment decisions in general practice. There is evidence that socioeconomic factors are associated with treatment decisions and treatment quality (29). Within the German national health insurance system, patients are able to receive permanent restorations without additional costs. A crown usually requires additional payment by the patient. Therefore, especially patients with lower income might tend to have their teeth restored with more extended restorations instead of having them crowned. In some of these cases, the indication for these extended restorations might be questionable or the individual tooth prognosis might be limited. These compromised treatment decisions might also negatively influence success rates. In other studies on re-interventions after restorative dental treatment, analyses were mainly conducted on a restoration level. Due to multiple methodological differences, great caution is required when comparing the results with ours. Some of these studies found similar or slightly superior success rates. On the tooth level, higher rates of re-interventions are expected compared to those on the restoration level. Data from the United Kingdom revealed success rates for posterior restorations between 60 % and 80 % at four years (10). The comparability of results however might allow drawing
the conclusion that the presented approach seems to be quite adequate to evaluate success and sustainability of dental restorations. Also based on massive data, Bogacki revealed success rates between 60 % and 94 % at four years for posterior restorations involving more than one surface (11) in the United States. Data from Scandinavia however show more comprehensive results. A study from Denmark involving 115 public health service dentists including 4,355 restorations revealed a success rate of 84 % at eight years (13). Although only focussing on specific restorations and patients, results from Finnish and Norwegian studies also describe superior outcomes (12, 28).
Especially when focussing on the sustainability aspect, our results raise questions. There are studies from single dental practices or practice networks available revealing better long-term success rates for permanent dental restorations (7, 8, 25, 27). Annual failure rates of 2 % (27) over 18 years or success rates of 65 % at 17 years (25) are published. To qualify these results, it must be clearly stated that results from selected practices or practice networks cannot be representative compared to general practice reality (30). The participation in research might affect a dentist’s treatment (9). However, these results indicate a certain potential for improvement within the health-care system.
Focussing on different restoration sizes and different tooth types, the results of this study are quite comprehensible. Focussing only on descriptive measures, two surface restorations showed highest success rates at four years while four and more surface restorations showed the lowest success rates. There are several studies describing an association between a increase in restoration size and lower success. On the face of it, the decreased performance of one surface restorations is surprising. However, these observed one surface restorations are not only comprising class I restorations but also class V restorations. Most studies describing a decreasing performance in relation to an expanded restoration size are only focusing on class I and class II restorations in premolars and molars (3, 10, 26) – or on primary teeth (31, 32). Having a comparable size, class V restorations tend to have lower
success rates (33) although comparing studies for different one surface restorations are rare (34). But there are also studies reporting favourable success rates of class V restorations (35). Additionally, teeth restored with a small one surface restoration might be under a higher risk for re-interventions on other surfaces. In conclusion, the wide range of one surface restorations - class I and class V - might be the main cause for their worse performance compared to two surface restorations. Surprisingly, the multivariate Cox regression analysis revealed a higher risk for re-interventions for two surface restorations compared to one surface restorations. This result underlines that not only the size of a restoration but also its tooth position and other parameters are important factors being associated with the risk for a re-intervention.
The presented study revealed significant differences of success rates between anterior teeth and posterior teeth. However, the differences are quite small, ranging between 0.4 % and 1.7 %. It has to be taken into account that large sample sizes reveal statistical significances although the magnitude of the differences itself may be minor and their clinical relevance questionable. The very small differences between anterior and posterior teeth are surprising because reasons for re-interventions can be quite different in anterior or posterior teeth. For example, poor appearance might be a major reason in anterior teeth. But that does not apply to posterior teeth. In that line, a recent review revealed different reasons for failure of restorations in anterior teeth (36) as they are known for posterior teeth (13, 14). In the presented study, small restorations (one surface) and very large restorations (more than three surfaces) perform slightly worse in anterior teeth compared to posterior teeth. For large restorations, this result seems logical. In anterior teeth, large restorations mostly comprise a part of the incisal edge which increases the risk for failure. One hypothesis to explain the very small differences in our study might be that re-interventions on a tooth level are mainly influenced by the level of destruction itself regardless of tooth types or positions. Respective comparable studies are lacking. Potential differences between primary and permanent teeth are expected but could not be derived from this study due to its study design.
Generalisation Due to different terms and conditions the presented results cannot be easily transferred to other health care systems. However, there are two significant reasons indicating a good external validity of this study. The first reason is that the German national health insurance system itself is quite non-restrictive. Patients and dentists can independently make their treatment decision while the insurance is covering the costs for standard treatment. In most cases, standard treatment means composite resin for class III, IV and V restorations and amalgam for class I and II restorations. If choosing a higher level of treatment the patient will have to cover the additional costs. The second reason is, that the level of dental treatment in Germany can be compared to other industrial countries worldwide. Therefore, the authors judge the range of the presented results transferable.
Conclusions Based on four year outcomes of more than seventeen million interventions, the sustainability of restorative dental treatment can be considered acceptable. However, from a health economics perspective, the results raise questions. As an overall goal, the need for restorative dental treatment should be decreased by strengthening primary and secondary prevention.
Acknowledgements The study was partially supported by the BARMER GEK – a German national health insurance company. The authors declare that there is no conflict of interest.
References 1. Sailer I, Makarov NA, Thoma DS, Zwahlen M, Pjetursson BE. All-ceramic or metalceramic tooth-supported fixed dental prostheses (FDPs)? A systematic review of the survival and complication rates. Part I: Single crowns (SCs). Dental Materials 2015; 31: 603-23. 2. Al-Khayatt AS, Ray-Chaudhuri A, Poyser NJ, Briggs PF, Porter RW, Kelleher MG, et al. Direct composite restorations for the worn mandibular anterior dentition: a 7-year followup of a prospective randomised controlled split-mouth clinical trial. Journal of Oral Rehabilitation 2013; 40: 389-401. 3. Opdam NJ, van de Sande FH, Bronkhorst E, Cenci MS, Bottenberg P, Pallesen U, et al. Longevity of posterior composite restorations: a systematic review and meta-analysis. Journal of Dental Research 2014; 93: 943-9. 4. Demarco FF, Correa MB, Cenci MS, Moraes RR, Opdam NJ. Longevity of posterior composite restorations: not only a matter of materials. Dental Materials 2012; 28: 87-101. 5. van Dijken JW, Pallesen U. A randomized 10-year prospective follow-up of Class II nanohybrid and conventional hybrid resin composite restorations. Journal of Adhesive Dentistry 2014; 16: 585-92. 6. Heintze SD, Rousson V. Clinical effectiveness of direct class II restorations - a metaanalysis. Journal of Adhesive Dentistry 2012; 14: 407-31. 7. Opdam NJ, Bronkhorst EM, Loomans BA, Huysmans MC. 12-year survival of composite vs. amalgam restorations. Journal of Dental Research 2010; 89: 1063-7. 8. Laske M, Opdam NJ, Bronkhorst EM, Braspenning JC, Huysmans MC. Longevity of direct restorations in Dutch dental practices. Descriptive study out of a practice based research network. Journal of Dentistry 2016; 46: 12-7. 9. Gilbert GH, Richman JS, Qvist V, Pihlstrom DJ, Foy PJ, Gordan VV. Change in stated clinical practice associated with participation in the Dental Practice-Based Research Network. General Dentistry 2010; 58: 520-8.
10. Burke FJ, Lucarotti PS. How long do direct restorations placed within the general dental services in England and Wales survive? British Dental Journal 2009; 206: E2; discussion 26-7. 11. Bogacki RE, Hunt RJ, del Aguila M, Smith WR. Survival analysis of posterior restorations using an insurance claims database. Operative Dentistry 2002; 27: 488-92. 12. Suni J, Vahanikkila H, Pakkila J, Tjaderhane L, Larmas M. Review of 36,537 patient records for tooth health and longevity of dental restorations. Caries Research 2013; 47: 309-17. 13. Pallesen U, van Dijken JW, Halken J, Hallonsten AL, Hoigaard R. A prospective 8-year follow-up of posterior resin composite restorations in permanent teeth of children and adolescents in Public Dental Health Service: reasons for replacement. Clinical Oral Investigations 2014; 18: 819-27. 14. Palotie U, Vehkalahti M. Reasons for replacement and the age of failed restorations in posterior teeth of young Finnish adults. Acta Odontologica Scandinavica 2002; 60: 325-9. 15. Rädel M, Hartmann A, Bohm S, Walter MH. Schriftenreihe zur Gesundheitsanalyse, Band 25 BARMER GEK Zahnreport 2014 Schwerpunkt: Wurzelkanalbehandlungen. . 1 ed. Siegburg: Asgard Verlagsservice; 2014. 16. Rädel M, Hartmann A, Bohm S, Walter MH. Schriftenreihe zur Gesundheitsanalyse, Band 31 BARMER GEK Zahnreport 2015 Schwerpunkt: Füllungen. 1 ed. Siegburg: Asgard Verlagsservice; 2015. 17. Peters MC, McLean ME. Minimally invasive operative care. I. Minimal intervention and concepts for minimally invasive cavity preparations. Journal of Adhesive Dentistry 2001; 3: 7-16. 18. Peters MC, McLean ME. Minimally invasive operative care. II. Contemporary techniques and materials: an overview. Journal of Adhesive Dentistry 2001; 3: 17-31. 19. Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal intervention dentistry--a review. FDI Commission Project 1-97. International Dental Journal 2000; 50: 1-12.
20. Burke FJ, Lucarotti PS. Ten-year outcome of crowns placed within the General Dental Services in England and Wales. Journal of Dentistry 2009; 37: 12-24. 21. Burke FJ, Lucarotti PS. Ten year survival of bridges placed in the General Dental Services in England and Wales. Journal of Dentistry 2012; 40: 886-95. 22. Raedel M, Hartmann A, Bohm S, Walter MH. Three-year outcomes of root canal treatment: Mining an insurance database. Journal of Dentistry 2015; 43: 412-7. 23. Raedel M, Hartmann A, Bohm S, Walter MH. Three-year outcomes of apicoectomy: mining an insurance database. Journal of Dentistry 2015. 24. Raedel M, Hartmann A, Bohm S, Konstantinidis I, Priess HW, Walter MH. Outcomes of direct pulp capping: interrogating an insurance database. International Endodontic Journal 2015. 25. da Rosa Rodolpho PA, Cenci MS, Donassollo TA, Loguercio AD, Demarco FF. A clinical evaluation of posterior composite restorations: 17-year findings. Journal of Dentistry 2006; 34: 427-35. 26. Gaengler P, Hoyer I, Montag R. Clinical evaluation of posterior composite restorations: the 10-year report. Journal of Adhesive Dentistry 2001; 3: 185-94. 27. van de Sande FH, Rodolpho PA, Basso GR, Patias R, da Rosa QF, Demarco FF, et al. 18-year survival of posterior composite resin restorations with and without glass ionomer cement as base. Dental Materials 2015; 31: 669-75. 28. Kopperud SE, Tveit AB, Gaarden T, Sandvik L, Espelid I. Longevity of posterior dental restorations and reasons for failure. European Journal of Oral Sciences 2012; 120: 53948. 29. Correa MB, Peres MA, Peres KG, Horta BL, Barros AJ, Demarco FF. Do socioeconomic determinants affect the quality of posterior dental restorations? A multilevel approach. Journal of Dentistry 2013; 41: 960-7. 30. Magin PJ, Marshall MJ, Goode SM, Cotter GL, Pond CD, Zwar NA. How generalisable are results of studies conducted in practice-based research networks? A cross-sectional
study of general practitioner demographics in two New South Wales networks. Medical Journal of Australia 2011; 195: 210-3. 31. da Franca C, Colares V, Van Amerongen E. Two-year evaluation of the atraumatic restorative treatment approach in primary molars class I and II restorations. International Journal of Paediatric Dentistry 2011; 21: 249-53. 32. Alves dos Santos MP, Luiz RR, Maia LC. Randomised trial of resin-based restorations in Class I and Class II beveled preparations in primary molars: 48-month results. Journal of Dentistry 2010; 38: 451-9. 33. Ceruti P, Menicucci G, Mariani GD, Pittoni D, Gassino G. Non carious cervical lesions. A review. Minerva Stomatologica 2006; 55: 43-57. 34. Sunico MC, Shinkai K, Katoh Y. Two-year clinical performance of occlusal and cervical giomer restorations. Operative Dentistry 2005; 30: 282-9. 35. Peumans M, De Munck J, Van Landuyt K, Van Meerbeek B. Thirteen-year randomized controlled clinical trial of a two-step self-etch adhesive in non-carious cervical lesions. Dental Materials 2015; 31: 308-14. 36. Demarco FF, Collares K, Coelho-de-Souza FH, Correa MB, Cenci MS, Moraes RR, et al. Anterior composite restorations: A systematic review on long-term survival and reasons for failure. Dental Materials 2015; 31: 1214-24.
Figure Legends
Figure 1
Patients over age groups (yrs = years)
Figure 2
Survival functions for the primary outcome re-intervention depending on the
size of the restoration. Figure 3
Survival functions for all target events (re-intervention, crowning, extraction)
depending on the size of the restoration.
Fig.1
Fig.2
Fig.3
Table 1
Four-year success rates for different restorations sizes in anterior and
posterior teeth (primary outcome: re-intervention). Restoration size /
Anterior teeth /
Posterior teeth /
involved tooth
Four-year success rate
Four-year success rate
surfaces
for the primary outcome
for the primary outcome
re-intervention in
re-intervention
p-Value
One
69.9 %
70.8 %
<0.0001
Two
74.3 %
73.7 %
<0.0001
Three
68.2 %
67.8 %
<0.0001
Four and more
62.1 %
63.8 %
<0.0001
Table 2
Cox regression analysis for all target events (re-intervention, crowning and
extraction) – detailed results Variable
p-Value
Hazard ratio Confidence interval
One surface restoration
reference
Two surface restoration
<0.0001
1.0594
1.0566 - 1.0621
Three surface restoration
<0.0001
1.0986
1.0954 - 1.1019
Four and more surface restoration
<0.0001
1.4409
1.4362 - 1.4456
Posterior tooth position
reference
Anterior tooth position
<0.0001
0.9993
0.9992 - 0.9994
S0300-5712(16)30243-3 http://dx.doi.org/doi:10.1016/j.jdent.2016.11.011 JJOD 2705
To appear in:
Journal of Dentistry
Received date: Revised date: Accepted date:
25-4-2016 23-10-2016 22-11-2016
Please cite this article as: Raedel Michael, Hartmann Andrea, Priess Heinz-Werner, Bohm Steffen, Samietz Stefanie, Konstantinidis Ioannis, Walter Michael H.Reinterventions after restoring teeth—Mining an insurance database.Journal of Dentistry http://dx.doi.org/10.1016/j.jdent.2016.11.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Re-interventions after restoring teeth – mining an insurance database
Short title:
Re-interventions after restoring teeth
Authors: Michael Raedela,
dentist /research associate [email protected]
Andrea Hartmannb,
biomedical statistician
[email protected]
Heinz-Werner Priessb biomedical statistician
[email protected]
Steffen Bohmb,
[email protected]
chief executive officer c
Stefanie Samietz
dentist /research associate [email protected]
Ioannis Konstantinidisa
dentist/research associate
[email protected] Michael H. Waltera,
professor and chair
[email protected]
Affiliations: a
Department of Prosthetic Dentistry
Technische Universität Dresden, Faculty of Medicine Carl Gustav Carus Fetscherstraße 74, 01307 Dresden, Germany b
AGENON
Gesellschaft für Forschung und Entwicklung im Gesundheitswesen Kaunstraße 21, 14163 Berlin, Germany c
Policlinic of Prosthetic Dentistry, Gerodontology and Biomaterials,
Center of Oral Health, University Medicine Greifswald, Rotgerberstr. 8, 17475 Greifswald, Germany
Corresponding author: Michael Raedel Department for Prosthetic Dentistry Faculty of Medicine Carl Gustav Carus Fetscherstraße 74 01307 Dresden Telephone:
+49 351 458-4231; Telefax: +49 351 458-5314
E-mail: [email protected]
Abstract Objectives: The aim of this study was to examine re-interventions after restorative treatment.
Methods: The data was collected from the digital database of a major German national health insurance company. Only permanent teeth were observed. Placing a permanent restoration other than a crown regardless of involved surfaces and material was the study intervention. The data did not allow for a differentiation between fillings and inlays that were estimated only a very small portion of the restorations. Success was defined as not undergoing any restorative re-intervention with fillings or inlays on the same tooth (primary outcome) and assessed with Kaplan-Meier survival analyses over four years. An additional analysis was conducted rating “crowning” and “extraction” of respective teeth as target events. Differences were tested with the Log-Rank-test. A multivariate Cox regression analyses was carried out.
Results: A total of 17,024,344 restorations placed in 4,825,408 anterior teeth and 9,973,177 posterior teeth could be traced. Focussing on the primary outcome re-intervention, the cumulative four-year success rate was 69.9 % for one surface restorations, 74.8 % for two surface restorations, 66.6 % for three surface restorations and 61.0 % for four surface and more extended restorations. These differences were significant (p<0.0001). Focussing on all three target events re-intervention, crowning and extraction, the cumulative four-year success rate was 66.1 % for one surface restorations, 67.5 % for two surface restorations, 63.0 % for three surface restorations and 55.8 % for four surface and more extended restorations. The number of restoration surfaces as well as the tooth position remained significant in the multivariate Cox regression.
Conclusions: The sustainability of restorative dental treatment under the terms and conditions of the German national health insurance system shows room for improvement.
From a public health perspective, special focus should be laid on primary and secondary prevention to minimize the restorative treatment need.
Clinical significance statement: This study shows that re-interventions are observed regularly after restorative treatment. Therefore, preventive and restorative strategies should be revisited and optimised.
key words:
dental restoration (permanent), general practice (dental), health services
research, operative dentistry, public health, treatment outcome
Introduction Caries is still one of the most prevalent diseases in dentistry worldwide. The unrecoverable loss of tooth structure caused by caries causes high costs for dental treatment in a long-term perspective. From the perspective of patients as well as from a public health perspective the sustainability of treatment is an important aspect in dentistry. There are treatment options that generally provide good sustainability because of their high long-term success or survival rates, for example single crowns (1). Regarding dental fillings and inlays, published success or survival rates differ. Especially when focussing on outcomes from general practice, data are rare.
Clinical studies and systematic reviews for different tooth types are available (2-6) mainly indicating favourable outcomes. However, it is quite obvious that results from clinical research are not representative for outcomes from general practice. There are results from single practices (7) and practice networks (8) delivering important data. Especially large practice based research networks (PBRN) using digital techniques for collecting data recently became an important addition to clinical research. Participating in research projects however changes attitudes and treatment decisions (9). Therefore, results from single practices or PBRN come closer to general practice but might still not be representative. At
the moment, the most reliable option for measuring outcomes directly and comprehensively from general practice can be realised in mining databases of public health care systems or public health insurances. These databases being originally created for payment of dentists, their scientific use may be often limited and methodologically restricted. However, these databases often provide treatment courses of several million patients.
Concerning restorative treatment, studies using massive data are quite rare. Highest case numbers were reached in studies conducted by Burke & Lucarotti using data from the British National Health Service (10, 11). In a four-year period, they found success rates roughly between 80 % for single surface amalgam restorations and 60 % for glass ionomer and composite resin restorations. A similar study based on private insurance data from the United States evaluated 300,753 patient cases (11). Success rates ranged between 60 % and 92 % at seven years. Other studies analysing treatment data from public health dental offices comprise much lower case numbers (13-14).
Therefore, the aim of this study was to extend the knowledge about the sustainability of restorative treatment under general practice conditions on the basis of a large data set. This data set consists of 8.6 million members of a German national health insurance company.
Materials and Methods The study was based on routine data of a major German national health insurance company (BARMER GEK, Berlin, Germany). This insurance company publishes annual health care reports (15, 16). In this context, the study group had access to the company’s data warehouse. The study design was approved by the responsible local ethics board (EK 288072015).
Fee codes and treatment dates for every single treatment step were available and allowed for tracing clinical courses on a day-count basis. Data were available for a four-year period from 01.01.2010 until 31.12.2013. Only data sets of patients that had been a member of the
insurance company for the whole four-year observation period entered the analysis. Because of systematically missing data, some specific German regions had to be excluded.
The placement of a permanent dental restoration other than a crown into a permanent tooth was defined as the initial study intervention. This marks the date of study intervention for the respective tooth. The claims data did not allow for a differentiation between fillings and inlays. The insurance system concedes different additional payments depending on the type of restoration that were not accessible for analysis. Inlays, however, were estimated only a very small portion of the restorations. Two independent survival analyses were calculated: In the first analysis, a re-intervention by placing a filling or an inlay on the respective tooth was defined as the primary outcome (primary re-intervention). This re-intervention can be a complete or partial renewal of the original restoration or a restoration independent from the original one. Within the primary outcome analysis, the occurrence of the events “crowning of the respective tooth” and “extraction of the respective tooth” were not rated as a target event but led to censoring. In a second analysis, all re-interventions were rated as target events. These re-interventions comprised the placing of a new filling or inlay in the respective tooth, the crowning of the respective tooth or the extraction of the respective tooth. Because permanent dental restorations may be used for building up teeth to be crowned, crowning was not counted as target event within 60 days after the restoration had been placed. The respective tooth type (anterior teeth, posterior teeth) was obtained from the database. The size of the dental restoration was categorized by the number of involved tooth surfaces (one, two, three, four and more extended). For all teeth that underwent no re-intervention, the observation period ended on December 31st 2013.
Survival analyses were conducted according to the Kaplan-Meier method. Differences between survival functions were first tested with the Log-Rank-Test. Additionally, a multivariate Cox-Regression analysis was carried out. The level of significance was set to p <
0.05. Survival analyses were conducted separately for each outcome on a day count basis. The software SAS (Statistical Analysis System, SAS Institute, Cary, NC, USA) was used for data preparation. The software R (available from http://www.r-project.org) with the add-on package “survival” was used for statistical analyses.
Results The study sample comprised 14,798,585 teeth with 17,024,344 restorations in 3,924,245 patients. The mean age was 51.4 years (Figure 1). A total of 4,825,408 anterior teeth and 9,973,177 posterior teeth were traced after having undergone a study intervention. The study intervention comprised 4,539,634 one surface restorations, 6,451,138 two surface restorations, 3,840,656 three surface restorations and 2,192,916 four surface and more extended restorations.
Focussing on the primary outcome re-intervention, one surface restorations showed a cumulative success rate of 69.9 % at four years. This value corresponds to an annual reintervention rate (ARR) of 7.5 %. Two surface, three surface and four surface and more extended restorations showed cumulative success rates of 74.8 % (ARR = 6.3 %), 66.6 % (ARR = 8.35 %) and 61.0 % (ARR = 9.75 %) at four years (Figure 2). The respective cumulative success rates at two years were 82.8 % (one surface), 85.2 % (two surface), 81.1 % (three surface) and 76.3 % (four surface and more extended). The Log-Rank test showed a significant difference between the survival functions (p<0.0001). For the secondary analysis comprising all three outcomes, the four-year success rates were 66.1 % for one surface restorations, 67.5 % for two surface restorations, 63.0 % for three surface restorations and 55.8 % for four surface and more extended restorations (Figure 3).
For the primary outcome, the mean observation period was 662.2 days for one surface restorations, 698.0 days for two surface restorations, 650.1 days for three surface
restorations and 622.4 days for four and more surface restorations. The respective median values were 638 days, 707 days, 616 days and 568 days. At two years, a number of 1,974,742 one surface restorations, 3,120,992 two surface restorations, 1,612,838 three surface restorations and 861,800 four surface and more extended restorations were still under observation. The respective numbers at the time of the last event were 5,821, 14,123, 9,394 and 10,065. The last events occurred between day 1,449 and day 1,456.
Comparing four-year success rates for equal restoration sizes in anterior and posterior teeth revealed small but significant differences. One surface restorations and four surface and more extended performed worse in anterior teeth compared to posterior teeth (Table 1). Two surface restorations and three surface restorations performed better in anterior teeth compared to posterior teeth.
In the multivariate Cox regression analysis, a higher number of restoration surfaces as well as a posterior tooth were found to be significantly associated with a higher risk for a reintervention (p<0.0001). Detailed results are shown in table 2.
Discussion Focussing on the primary outcome re-intervention, the presented study revealed cumulative success rates between 60 % and 70 % at four years based on an exceptional high number of cases. However, some limitations have to be taken into account in order to facilitate an appropriate discussion and to avoid over-interpretation.
Limitation The major limitation of this study was the data source being not a clinical documentation but fee code based data provided by an insurance company. On the one hand, the database comprised the complete number of all conducted treatments for the 8.6 million members of
this national health insurance company representing 10.5 % of the German population in December 2013. On the other hand, there was no additional clinical information except for the documented treatment fee codes which are digitally submitted from the dentist to the insurance company without involving the patient. However, given appropriate interpretation, the value of the respective results is unquestionable. Health Services Research and the use of massive data resources for evaluating outcomes from general dental practice are gaining increasing importance within the scientific community. In this context, the aspect of data quality has to be mentioned. Fee code based data may be compromised by a mix-up of teeth, surfaces and fee codes although the respective errors may to a certain extent compensate for one another. However, a potential bias has to be taken into account but is all in all judged minor.
Dental restorations were detected based on respective treatment fee codes. These codes did not deliver information about materials (composite resins, amalgam, glass ionomer cements, different inlay materials and other) or techniques. An estimation of the distribution of the material for posterior teeth can be derived from the results of a limited explorative survey conducted by the study group independently from this analysis. Within this survey, patients were asked about the materials used for restorations in their first molars. Composite materials showed a percentage of 44 % followed by amalgam with 20 %. 21 % of the patients did not remember. All other materials were below 10 %. Comparable studies based on data from England and Wales were able to detect differences between materials (10). The lack of respective data is a second major limitation.
Re-interventions were evaluated on the tooth level. This unusual approach was chosen with respect to the sustainability of restorative dental treatment. Of course, the presented results will somehow overestimate re-intervention rates on a restoration level. Furthermore, modern concepts for restoring teeth even prefer small restorations and repairs over larger restorations in order to save tooth structure (17-19). However, from the patients’ and the
health economics perspectives, the sustainability of restorative treatment is an important issue. After placement of a restoration other than a crown, a tooth should be seen as completely restored in a medium-term perspective. For evaluating the sustainability of the treatment concept, this study aimed at detecting each re-intervention on the same tooth within the four year period. This is feasible although the re-intervention might not directly affect the initial restoration. Due to this methodology, the presented results cannot be directly compared with restoration based success or survival rates.
The placement of a new restoration other than a crown at the same tooth was chosen as the primary outcome. It can be seen as the most likely event associated with the previous placement of a dental restoration. Crowning or extracting a tooth might also be associated with the previous placement of a dental restoration but several other reasons come into question. Crowning and extraction therefore served as secondary outcomes. Due to these different indications, a combined analysis was considered inappropriate. The four-year observation period was limited by the availability of the data. It can be judged acceptable to evaluate the sustainability in a medium term perspective. Regarding the statistical methods, the Kaplan-Meier survival analysis can be seen as a common method for describing outcomes after interventions in medical and dental studies. It is an accepted method for outcome analyses on the basis of massive data (10, 11, 20-24).
Interpretation Interpreting the presented results is difficult due to the lack of comparable studies. In this dimension, there is no study available focussing on the outcome of anterior and posterior teeth on a tooth level. The vast majority of studies or reviews only focus on posterior teeth (37, 10, 11, 13, 14, 25-28). Nevertheless, we see two major aspects for interpretation. One aspect is the range of the presented success rates itself in comparison to success/survival rates from other studies. The other aspect can be seen in the significant differences this study revealed between success/survival rates for different tooth types and restoration sizes.
Focussing on the primary outcome re-intervention, the overall success rates at four years were roughly found between 60 % and 75 %. That is, at least every 4th tooth underwent a reintervention within four years after a permanent restoration. Results from practice based research networks tend to be superior (8). Furthermore, taking all outcomes into account, the results raise even more questions. But the presented results are not representing restoration based success or survival rates. The performance of a single restoration might be underestimated by this approach. Therefore, the overall judgement is difficult and somehow controversial.
There are several reasons for considering our results to be acceptable. The major argument might be seen in the contemporary preference of restoring defects as small as possible (1719). This might tend to a higher number of interventions itself. A second point is that socioeconomic factors might affect treatment decisions in general practice. There is evidence that socioeconomic factors are associated with treatment decisions and treatment quality (29). Within the German national health insurance system, patients are able to receive permanent restorations without additional costs. A crown usually requires additional payment by the patient. Therefore, especially patients with lower income might tend to have their teeth restored with more extended restorations instead of having them crowned. In some of these cases, the indication for these extended restorations might be questionable or the individual tooth prognosis might be limited. These compromised treatment decisions might also negatively influence success rates. In other studies on re-interventions after restorative dental treatment, analyses were mainly conducted on a restoration level. Due to multiple methodological differences, great caution is required when comparing the results with ours. Some of these studies found similar or slightly superior success rates. On the tooth level, higher rates of re-interventions are expected compared to those on the restoration level. Data from the United Kingdom revealed success rates for posterior restorations between 60 % and 80 % at four years (10). The comparability of results however might allow drawing
the conclusion that the presented approach seems to be quite adequate to evaluate success and sustainability of dental restorations. Also based on massive data, Bogacki revealed success rates between 60 % and 94 % at four years for posterior restorations involving more than one surface (11) in the United States. Data from Scandinavia however show more comprehensive results. A study from Denmark involving 115 public health service dentists including 4,355 restorations revealed a success rate of 84 % at eight years (13). Although only focussing on specific restorations and patients, results from Finnish and Norwegian studies also describe superior outcomes (12, 28).
Especially when focussing on the sustainability aspect, our results raise questions. There are studies from single dental practices or practice networks available revealing better long-term success rates for permanent dental restorations (7, 8, 25, 27). Annual failure rates of 2 % (27) over 18 years or success rates of 65 % at 17 years (25) are published. To qualify these results, it must be clearly stated that results from selected practices or practice networks cannot be representative compared to general practice reality (30). The participation in research might affect a dentist’s treatment (9). However, these results indicate a certain potential for improvement within the health-care system.
Focussing on different restoration sizes and different tooth types, the results of this study are quite comprehensible. Focussing only on descriptive measures, two surface restorations showed highest success rates at four years while four and more surface restorations showed the lowest success rates. There are several studies describing an association between a increase in restoration size and lower success. On the face of it, the decreased performance of one surface restorations is surprising. However, these observed one surface restorations are not only comprising class I restorations but also class V restorations. Most studies describing a decreasing performance in relation to an expanded restoration size are only focusing on class I and class II restorations in premolars and molars (3, 10, 26) – or on primary teeth (31, 32). Having a comparable size, class V restorations tend to have lower
success rates (33) although comparing studies for different one surface restorations are rare (34). But there are also studies reporting favourable success rates of class V restorations (35). Additionally, teeth restored with a small one surface restoration might be under a higher risk for re-interventions on other surfaces. In conclusion, the wide range of one surface restorations - class I and class V - might be the main cause for their worse performance compared to two surface restorations. Surprisingly, the multivariate Cox regression analysis revealed a higher risk for re-interventions for two surface restorations compared to one surface restorations. This result underlines that not only the size of a restoration but also its tooth position and other parameters are important factors being associated with the risk for a re-intervention.
The presented study revealed significant differences of success rates between anterior teeth and posterior teeth. However, the differences are quite small, ranging between 0.4 % and 1.7 %. It has to be taken into account that large sample sizes reveal statistical significances although the magnitude of the differences itself may be minor and their clinical relevance questionable. The very small differences between anterior and posterior teeth are surprising because reasons for re-interventions can be quite different in anterior or posterior teeth. For example, poor appearance might be a major reason in anterior teeth. But that does not apply to posterior teeth. In that line, a recent review revealed different reasons for failure of restorations in anterior teeth (36) as they are known for posterior teeth (13, 14). In the presented study, small restorations (one surface) and very large restorations (more than three surfaces) perform slightly worse in anterior teeth compared to posterior teeth. For large restorations, this result seems logical. In anterior teeth, large restorations mostly comprise a part of the incisal edge which increases the risk for failure. One hypothesis to explain the very small differences in our study might be that re-interventions on a tooth level are mainly influenced by the level of destruction itself regardless of tooth types or positions. Respective comparable studies are lacking. Potential differences between primary and permanent teeth are expected but could not be derived from this study due to its study design.
Generalisation Due to different terms and conditions the presented results cannot be easily transferred to other health care systems. However, there are two significant reasons indicating a good external validity of this study. The first reason is that the German national health insurance system itself is quite non-restrictive. Patients and dentists can independently make their treatment decision while the insurance is covering the costs for standard treatment. In most cases, standard treatment means composite resin for class III, IV and V restorations and amalgam for class I and II restorations. If choosing a higher level of treatment the patient will have to cover the additional costs. The second reason is, that the level of dental treatment in Germany can be compared to other industrial countries worldwide. Therefore, the authors judge the range of the presented results transferable.
Conclusions Based on four year outcomes of more than seventeen million interventions, the sustainability of restorative dental treatment can be considered acceptable. However, from a health economics perspective, the results raise questions. As an overall goal, the need for restorative dental treatment should be decreased by strengthening primary and secondary prevention.
Acknowledgements The study was partially supported by the BARMER GEK – a German national health insurance company. The authors declare that there is no conflict of interest.
References 1. Sailer I, Makarov NA, Thoma DS, Zwahlen M, Pjetursson BE. All-ceramic or metalceramic tooth-supported fixed dental prostheses (FDPs)? A systematic review of the survival and complication rates. Part I: Single crowns (SCs). Dental Materials 2015; 31: 603-23. 2. Al-Khayatt AS, Ray-Chaudhuri A, Poyser NJ, Briggs PF, Porter RW, Kelleher MG, et al. Direct composite restorations for the worn mandibular anterior dentition: a 7-year followup of a prospective randomised controlled split-mouth clinical trial. Journal of Oral Rehabilitation 2013; 40: 389-401. 3. Opdam NJ, van de Sande FH, Bronkhorst E, Cenci MS, Bottenberg P, Pallesen U, et al. Longevity of posterior composite restorations: a systematic review and meta-analysis. Journal of Dental Research 2014; 93: 943-9. 4. Demarco FF, Correa MB, Cenci MS, Moraes RR, Opdam NJ. Longevity of posterior composite restorations: not only a matter of materials. Dental Materials 2012; 28: 87-101. 5. van Dijken JW, Pallesen U. A randomized 10-year prospective follow-up of Class II nanohybrid and conventional hybrid resin composite restorations. Journal of Adhesive Dentistry 2014; 16: 585-92. 6. Heintze SD, Rousson V. Clinical effectiveness of direct class II restorations - a metaanalysis. Journal of Adhesive Dentistry 2012; 14: 407-31. 7. Opdam NJ, Bronkhorst EM, Loomans BA, Huysmans MC. 12-year survival of composite vs. amalgam restorations. Journal of Dental Research 2010; 89: 1063-7. 8. Laske M, Opdam NJ, Bronkhorst EM, Braspenning JC, Huysmans MC. Longevity of direct restorations in Dutch dental practices. Descriptive study out of a practice based research network. Journal of Dentistry 2016; 46: 12-7. 9. Gilbert GH, Richman JS, Qvist V, Pihlstrom DJ, Foy PJ, Gordan VV. Change in stated clinical practice associated with participation in the Dental Practice-Based Research Network. General Dentistry 2010; 58: 520-8.
10. Burke FJ, Lucarotti PS. How long do direct restorations placed within the general dental services in England and Wales survive? British Dental Journal 2009; 206: E2; discussion 26-7. 11. Bogacki RE, Hunt RJ, del Aguila M, Smith WR. Survival analysis of posterior restorations using an insurance claims database. Operative Dentistry 2002; 27: 488-92. 12. Suni J, Vahanikkila H, Pakkila J, Tjaderhane L, Larmas M. Review of 36,537 patient records for tooth health and longevity of dental restorations. Caries Research 2013; 47: 309-17. 13. Pallesen U, van Dijken JW, Halken J, Hallonsten AL, Hoigaard R. A prospective 8-year follow-up of posterior resin composite restorations in permanent teeth of children and adolescents in Public Dental Health Service: reasons for replacement. Clinical Oral Investigations 2014; 18: 819-27. 14. Palotie U, Vehkalahti M. Reasons for replacement and the age of failed restorations in posterior teeth of young Finnish adults. Acta Odontologica Scandinavica 2002; 60: 325-9. 15. Rädel M, Hartmann A, Bohm S, Walter MH. Schriftenreihe zur Gesundheitsanalyse, Band 25 BARMER GEK Zahnreport 2014 Schwerpunkt: Wurzelkanalbehandlungen. . 1 ed. Siegburg: Asgard Verlagsservice; 2014. 16. Rädel M, Hartmann A, Bohm S, Walter MH. Schriftenreihe zur Gesundheitsanalyse, Band 31 BARMER GEK Zahnreport 2015 Schwerpunkt: Füllungen. 1 ed. Siegburg: Asgard Verlagsservice; 2015. 17. Peters MC, McLean ME. Minimally invasive operative care. I. Minimal intervention and concepts for minimally invasive cavity preparations. Journal of Adhesive Dentistry 2001; 3: 7-16. 18. Peters MC, McLean ME. Minimally invasive operative care. II. Contemporary techniques and materials: an overview. Journal of Adhesive Dentistry 2001; 3: 17-31. 19. Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal intervention dentistry--a review. FDI Commission Project 1-97. International Dental Journal 2000; 50: 1-12.
20. Burke FJ, Lucarotti PS. Ten-year outcome of crowns placed within the General Dental Services in England and Wales. Journal of Dentistry 2009; 37: 12-24. 21. Burke FJ, Lucarotti PS. Ten year survival of bridges placed in the General Dental Services in England and Wales. Journal of Dentistry 2012; 40: 886-95. 22. Raedel M, Hartmann A, Bohm S, Walter MH. Three-year outcomes of root canal treatment: Mining an insurance database. Journal of Dentistry 2015; 43: 412-7. 23. Raedel M, Hartmann A, Bohm S, Walter MH. Three-year outcomes of apicoectomy: mining an insurance database. Journal of Dentistry 2015. 24. Raedel M, Hartmann A, Bohm S, Konstantinidis I, Priess HW, Walter MH. Outcomes of direct pulp capping: interrogating an insurance database. International Endodontic Journal 2015. 25. da Rosa Rodolpho PA, Cenci MS, Donassollo TA, Loguercio AD, Demarco FF. A clinical evaluation of posterior composite restorations: 17-year findings. Journal of Dentistry 2006; 34: 427-35. 26. Gaengler P, Hoyer I, Montag R. Clinical evaluation of posterior composite restorations: the 10-year report. Journal of Adhesive Dentistry 2001; 3: 185-94. 27. van de Sande FH, Rodolpho PA, Basso GR, Patias R, da Rosa QF, Demarco FF, et al. 18-year survival of posterior composite resin restorations with and without glass ionomer cement as base. Dental Materials 2015; 31: 669-75. 28. Kopperud SE, Tveit AB, Gaarden T, Sandvik L, Espelid I. Longevity of posterior dental restorations and reasons for failure. European Journal of Oral Sciences 2012; 120: 53948. 29. Correa MB, Peres MA, Peres KG, Horta BL, Barros AJ, Demarco FF. Do socioeconomic determinants affect the quality of posterior dental restorations? A multilevel approach. Journal of Dentistry 2013; 41: 960-7. 30. Magin PJ, Marshall MJ, Goode SM, Cotter GL, Pond CD, Zwar NA. How generalisable are results of studies conducted in practice-based research networks? A cross-sectional
study of general practitioner demographics in two New South Wales networks. Medical Journal of Australia 2011; 195: 210-3. 31. da Franca C, Colares V, Van Amerongen E. Two-year evaluation of the atraumatic restorative treatment approach in primary molars class I and II restorations. International Journal of Paediatric Dentistry 2011; 21: 249-53. 32. Alves dos Santos MP, Luiz RR, Maia LC. Randomised trial of resin-based restorations in Class I and Class II beveled preparations in primary molars: 48-month results. Journal of Dentistry 2010; 38: 451-9. 33. Ceruti P, Menicucci G, Mariani GD, Pittoni D, Gassino G. Non carious cervical lesions. A review. Minerva Stomatologica 2006; 55: 43-57. 34. Sunico MC, Shinkai K, Katoh Y. Two-year clinical performance of occlusal and cervical giomer restorations. Operative Dentistry 2005; 30: 282-9. 35. Peumans M, De Munck J, Van Landuyt K, Van Meerbeek B. Thirteen-year randomized controlled clinical trial of a two-step self-etch adhesive in non-carious cervical lesions. Dental Materials 2015; 31: 308-14. 36. Demarco FF, Collares K, Coelho-de-Souza FH, Correa MB, Cenci MS, Moraes RR, et al. Anterior composite restorations: A systematic review on long-term survival and reasons for failure. Dental Materials 2015; 31: 1214-24.
Figure Legends
Figure 1
Patients over age groups (yrs = years)
Figure 2
Survival functions for the primary outcome re-intervention depending on the
size of the restoration. Figure 3
Survival functions for all target events (re-intervention, crowning, extraction)
depending on the size of the restoration.
Fig.1
Fig.2
Fig.3
Table 1
Four-year success rates for different restorations sizes in anterior and
posterior teeth (primary outcome: re-intervention). Restoration size /
Anterior teeth /
Posterior teeth /
involved tooth
Four-year success rate
Four-year success rate
surfaces
for the primary outcome
for the primary outcome
re-intervention in
re-intervention
p-Value
One
69.9 %
70.8 %
<0.0001
Two
74.3 %
73.7 %
<0.0001
Three
68.2 %
67.8 %
<0.0001
Four and more
62.1 %
63.8 %
<0.0001
Table 2
Cox regression analysis for all target events (re-intervention, crowning and
extraction) – detailed results Variable
p-Value
Hazard ratio Confidence interval
One surface restoration
reference
Two surface restoration
<0.0001
1.0594
1.0566 - 1.0621
Three surface restoration
<0.0001
1.0986
1.0954 - 1.1019
Four and more surface restoration
<0.0001
1.4409
1.4362 - 1.4456
Posterior tooth position
reference
Anterior tooth position
<0.0001
0.9993
0.9992 - 0.9994