Several new locally administered drug treatments that are less invasive and perhaps less expensive than surgery have been developed. Consideration of these modalities may be important because treatment costs for surgical services generally are higher than those for nonsurgical services. Indeed, any new treatments need to be assessed in terms of their economic impact as well as in terms of their clinical value.

The chlorhexidine, or CHX, chip (PerioChip 2.5 milligrams [chlorhexidine gluconate], manufactured by Perio Products Ltd., Jerusalem; distributed by Dexcel Pharma, Edison, N.J.) is one of these new treatments. The CHX chip is a small, biodegradable chip of hydrolyzed gelatin containing the antimicrobial chlorhexidine gluconate. It is placed directly in periodontal pockets and releases CHX at the site for seven to 10 days.³ Clinical studies have shown that adjunctive use of the CHX chip in patients with chronic periodontitis reduces PPD, improves clinical attachment levels and reduces bleeding on probing compared with SRP alone.^4,5

A previous report modeled the cost-effectiveness of SRP augmented by the CHX chip, or SRP + CHX chip, compared with SRP alone, based on clinical trial data⁴ and expert opinion.⁶ The data described patient periodontal status (that is, PPD and bleeding on probing) after treatment, and periodontal experts (two of the trial investigators) estimated the proportion of patients who might require surgery as a function of their PPD and bleeding-on-probing status. The model predicted reduced surgical needs for patients treated with SRP + CHX chip compared with SRP alone for little additional cost.⁶ Now we have tested these conclusions in actual clinical practice, conducting a prospective clinical trial to assess the economic impact of the CHX chip in general dental practice. Our article reports the results of this cost comparison.

The primary objective of this economic study was to compare the charges for treating chronic periodontitis with SRP vs. SRP + CHX chip. The secondary objectives were to determine the effect of the CHX chip on periodontal procedures (such as periodontal surgery and tooth extractions) performed over 12 months in general dental practices and to compare treatment recommendations by periodontal specialists after patients had completed one year under each treatment.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Study design. This study was a multicenter, single-blind, randomized, parallel group clinical trial in which we compared patients who had chronic periodontitis and received SRP + CHX chip with patients who received only SRP. The intent of the study was to assess the treatments that periodontitis patients received under the care of general dental practitioners over a period of one year, as well as the costs associated with those treatments. Patients in the study were enrolled in 52 general dental practices across the United States. After enrollment, patients were evaluated by periodontists in 10 U.S. academic evaluation centers (among them, R.J.G. and W.J.K.) to confirm their eligibility for the study before random treatment assignment. Patients were randomized at the general practices into two treatment arms: SRP or SRP + CHX chip. The general dentists who were administering the regular care were aware of each patient’s treatment assignment. The participating academic periodontist evaluators were blinded to the randomization throughout the study.

Patients were seen by general dentists at mandatory office visits at baseline and at three, six, nine and 12 months. Three-month recall visits typically included a periodontal prophylaxis, as appropriate. Assessments by the academic evaluators occurred at baseline and at six and 12 months, just before the respective scheduled office visits; assessments included PPD and probing attachment level, or PAL. At the baseline and 12-month visits, the academic evaluators recommended treatments that they thought were warranted at those times, respectively. The recommendations were not divulged to the treating general dentists. A schematic of the study time-line is pictured in Figure 1.

View larger version (29K): [in this window] [in a new window]   Figure 1. The study timeline. The authors conducted a one-year prospective clinical trial in which patients underwent one of two treatment regimens: either scaling and root planing, or SRP, plus any therapy deemed necessary by unblinded, treating general dentists; or SRP plus other therapy as above plus adjunctive administration of a chlorhexidine chip. Blinded periodontists also evaluated patients at academic evaluation centers.

Study patients. The patient group included both men and women. Eligibility criteria included the following: age of 30 to 65 years, inclusive; good general health; at least 20 natural teeth; and presence of periodontitis characterized by at least two quadrants with disease, each of which had at least two teeth that each had one or more pockets with PPD 5 mm. Exclusion criteria included the following: having any pocket with PPD 10 mm; having undergone active periodontal therapy within the year preceding study enrollment; concomitantly taking medications known to affect gingival or periodontal tissues; and having severe periodontal disease characterized by bone loss to within 3 mm of the apex of any tooth. The study cohort for analysis consisted of all randomized patients who completed the study through nine months, defined to be anyone who completed visits at months nine or 12 at the general dental office regardless of missed prior visits. The sample size of 484 was based on expected rates of undergoing surgery within one year: 83 percent and 69 percent in the SRP group and the SRP + CHX chip group, respectively. We made these estimates assuming that treatment in actual practice would not differ substantially from expert recommendations for surgery that we had modeled previously.⁶

Perspective. The primary perspective of this analysis was that of a comprehensive third-party dental payer providing full coverage. The billed charges for all services were the primary economic measure. We included in the calculations all services that might be covered by a third-party dental payer.

Charge and utilization data. We collected economic data in three ways: asking study centers to provide bills for all services that occurred at the dental offices and after referral; creating case report forms that included for each study visit a page on which to report referrals that were made (and completed) in the prior period and whether or not a bill was collected for each referral visit; stipulating that blinded evaluators recommend services that they thought were warranted.

All services reported in bills and recommended by evaluators were coded by a trained dental coder using Current Dental Terminology-2, or CDT-2, codes.⁷ Study centers were asked to collect bills for all dental services that occurred during the 12 months of the study. However, for each visit for which a bill was missing, the study monitors contacted dental offices to determine the services that were provided, which the dental coder then coded. We then coded and priced provided services at the average charge for services with the respective CDT-2 codes in our sample to simulate the missing charges. One author (R.D.F.), a periodontist who was blinded to treatment assignments, categorized all services into major service groups (such as periodontics and prosthodontics) using CDT-2 codes before analysis took place.

The coder classified the services recommended by evaluation centers according to CDT-2 codes and priced them using the mean fees from the 1995 American Dental Association Survey of Dental Fees⁸ (inflated to 1997 levels using the dental care services index from the medical care component of the consumer price index, for an inflation factor of 1.095⁹). This method provided standard prices for services. When procedures or services were not included in the ADA survey, we estimated the values for them by calculating the average ratio of ADA survey fees to charges from the bills collected in this study over all services contained in both the ADA survey and the study-derived charge database. In this average, each service received equal weight (that is, we calculated the ratio for each procedure, as well as a simple mean of these procedure-specific ratios). We calculated only a single ratio, using data for all specialties and geographic regions as they were contained in the charge database. We calculated the simulated ADA survey values for codes not contained in the survey by multiplying the average of the reported charges for each service by this overall ratio.

We conducted a check for unit charge (price) outliers. Charges at least five standard deviations higher than the mean for individual services were replaced with the mean charge for that service. During the vast majority of the clinical trial (patients were seen in the study from Oct. 23, 1996, to June 9, 1998), the CHX chip was an investigational drug in the United States and not yet cleared for distribution. The CHX chip was provided at no charge to the study investigators and used at no charge for the study patients. For the purpose of assigning a cost for the CHX chip for study analysis, since the average wholesale price typically is used as the drug cost for economic analyses¹⁰ and the CHX chip price was projected to vary by the quantity ordered, we used the projected midrange wholesale price of $14 per chip.

We defined total dental charges as the total charge for all dental services used during the 12 months of the study (including the services for which bills were not available). For the SRP + CHX chip group, we calculated these total charges with and without the charges for the CHX chip. Several subsets of these charges are reported as well, with particular attention given to total periodontal charges. We estimated charges for services recommended at the final evaluation by pricing the services recommended (as described above) and summing across all services.

Statistical methods. We conducted both univariate and multivariate analyses. Univariate analyses are presented in addition to multivariate analyses for the utilization measures, because the probabilities obtained from the univariate analyses are more natural measures than the odds ratios estimated in the multivariate analyses of these binary outcomes.

Because of differences in baseline severity between the two study groups, we considered the multivariate results (the adjusted analyses) to provide more robust population estimates than the univariate analyses. Therefore, we report only the adjusted analyses for charges. The multivariate models used for charges were linear regressions of each specific charge measure on fixed effects for baseline parameters, including treatment group, sex, age group, race, smoking status, PAL severity classification, PPD classification, number of teeth with at least one pocket with PPD 5 mm and number of quadrants containing at least one tooth that had at least one pocket with PPD 5 mm. We used random effects for region and specific study site. The adjusted charges for each study group were least squares means, and the significance levels were based on an F test of the fixed effect for study group. The multivariate models for utilization measures were logistic regression equations with the same fixed effects as the charge models, but no random effects. We used the simpler model specification for the logistic regression model because of the scarcity of events. Statistical significance was defined as P .05. We conducted a sensitivity analysis for total dental charges by using unit charges taken from the ADA survey rather than prices actually charged by dentists at each practice.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Patient characteristics. A total of 454 of 484 randomized patients completed the study through nine months, regardless of missed prior visits, and 451 completed it through 12 months. Ninety-four percent of patients completed the study in both treatment arms. The 6 percent who did not qualify for the study cohort differed from those who did qualify (P .05) in number of teeth having at least one pocket with PPD 5 mm and number of quadrants containing at least one tooth that had at least one pocket with PAL 5 mm. They did not differ in demographics, smoking history or other baseline severity measures.

The SRP group and the SRP + CHX chip group had similar demographics and medical and dental histories (Table 1). While gross measures of periodontal status (PAL severity classification, PPD severity classification, number of natural teeth) did not differ between the groups, the SRP + CHX chip group had a significantly larger number of both quadrants containing at least one tooth that had at least one site with PAL 5 mm and teeth that had at least one site with PAL 5 mm. In addition, the maximum PPD was numerically higher in the SRP + CHX chip group, although not significantly so. These differences, which occurred by chance despite the study randomization process, implied that analyses adjusted for baseline characteristics were the most appropriate for comparing the two treatment approaches.

Dental care charges. Table 4 (page 1564) shows the adjusted charges per patient for dental services rendered throughout 12 months. Total dental charges including CHX chip costs were significantly higher for the SRP + CHX chip group compared with those for the SRP group (P = .027). When CHX chip costs were excluded, SRP + CHX chip was less expensive than SRP alone (P = .012). The magnitude of the difference was similar (but opposite in sign) between these two charge measures, suggesting that about one-half of the acquisition cost of the CHX chip was offset by savings in other dental charges during the 12-month study. The two treatment groups did not differ in charges for diagnostic, preventive or restorative services; endodontic services; oral and maxillofacial surgery; orthodontic services; adjunctive general services, total periodontal services, or SRP alone. However, SRP + CHX chip group charges were significantly lower than SRP group charges for periodontal surgery (P = .016) and for prosthodontics (P = .016). The estimated charge differences in these two categories of services accounted for more than one-half of the difference in total charges without the CHX chip.

View larger version (37K): [in this window] [in a new window]   Figure 2. Total dental charges assessed in the study, by quarter. SRP: Scaling and root planing. CHX chip: Chlorhexidine chip.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Primary findings. There were several notable findings in this study.

Cost comparison. Firstly, the primary objective of this study was to compare costs of one year’s therapy in general practice for patients with chronic periodontitis who were treated with or without the CHX chip. Costs for services for patients treated adjunctively with the CHX chip were significantly greater than for those treated without it. Treatment with the CHX chip lowered other charges enough to offset about one-half of the chip acquisition costs. When the charges for the chip were included, the overall charges per patient for the SRP + CHX chip group were $175 higher than those for the SRP group.

Likelihood of undergoing surgery. Secondly, patients in the SRP + CHX chip group had only about half as much likelihood of receiving surgical services as did those in the SRP group. Being able to control disease nonsurgically has cost implications, as well as quality-of-life implications. With respect to the cost analysis, the incremental increase in cost resulted in an incremental reduction in the percentage of patients treated with surgery. Patients may be quite willing to support a modest increase in costs to avoid surgery.

Differences in care received vs. care recommended. Thirdly, there were clear differences between the care that patients actually received during the 12-month study period and the care that was recommended at 12 months by the periodontal evaluators. For example, total average dental charges per patient for the services rendered were $1,393 and $1,568 for the SRP group and the SRP + CHX chip group, respectively, but the total dental charges for the services recommended by the periodontists at the final evaluation were $2,097 and $2,119 for the SRP group and the SRP + CHX chip group, respectively. Even more striking is the fact that the average charge per patient for periodontal surgery actually rendered was $46 for the SRP group and $17 for the SRP + CHX chip group, but the average charge per patient for periodontal surgery that was recommended by the academic evaluators was $1,343 for the SRP group and $1,392 for the SRP + CHX chip group. Also notably, the average charge per patient for prosthodontics was $118 for the SRP group and $23 for the SRP + CHX chip group, while the average charge per patient for prosthodontics and implants that were recommended by the evaluators was $270 for the SRP group and $239 for the SRP + CHX chip group. The dental office data reflect actual treatment rendered over 12 months of therapy, which included a six-month period for surgical therapy, while the evaluation center recommendations represent study-imposed prospective assessments by periodontal specialists at the end of the trial. Finally, the majority of the cost of treatment for the 12-month period occurred in the first three months of the study.

Issues of interpretation. A number of questions may be raised about the interpretation of these study data. First of all, the general dentists performing treatment in this study were not blinded as to patient randomization; they were aware of whether patients were being treated with or without the CHX chip. Thus, one could question whether the resultant reduction of surgery for patients in the SRP + CHX chip group was due to a systematic bias because of this lack of blinding. Since the general dentists were participating in a study to address the potential of the CHX chip as a nonsurgical treatment, they may have systematically performed or recommended fewer surgeries for patients in the SRP + CHX chip group. The lack of blinding clearly could have influenced both hygienists’ recommendations and patient preferences as well. The data reflect only what treatments were performed in this study regardless of any bias. In actual practice, however, neither practitioners nor patients are blinded to treatments. Thus, systematic biases in this study actually may reflect biases that normally could exist in clinical practice.

Differences in amount of surgery performed. An important question about the interpretation of the study results regards the difference between the amount of surgery performed during the year for the patients of the participating general practitioners and the amount of surgery recommended by periodontal evaluators at the end of the study. It is clear that there was a systematic bias for both treatment groups to have far less surgery than was either recommended by periodontists at the end of the study or anticipated before the start of the study, based on our previously modeled predictions.⁶ While less than 16 percent of patients underwent surgery during the study (15.5 percent, SRP group; 9.2 percent, SRP + CHX chip group), more than 55 percent of patients were recommended for surgery by periodontists at the end of the study (55.9 percent, SRP group; 62.9 percent, SRP + CHX chip group).

The data suggest that general practitioners had placed their respective patients predominantly into maintenance programs and were treating with supportive periodontal therapy at three-month intervals. However, the data do not make any suggestion about the interpretation of the difference between the treatment performed (and presumably recommended) during the study by general practitioners and the treatment recommended by specialists at the end of the study. Although it is not known how much periodontal surgery might have been recommended by the general dentists in this study but not performed, it seems unlikely that this amount would have equaled that recommended by the periodontists. On the one hand, both general dentists and patients knew they were involved in a study of nonsurgical therapy, which could have influenced treatment decisions. The "placebo effect" in studies, often but ambiguously termed the "Hawthorne effect," is well-known.¹¹ It also could be speculated that the general dentists, who had less training in the treatment of periodontal diseases, "undertreated" their patients and provided fewer procedures (for instance, surgery) than needed, needs which were identified by specialists at 12 months. As suggested by the periodontal recommendations at the end of the study, if the general dentists did not adequately treat the periodontal disease in their patients, presumably these patients then were at risk of experiencing increased disease progression, which then could result in increased long-term dental needs and charges.

On the other hand, one could also speculate that because of the study design, the general dentists in fact knew that they were being watched, that their patients (and their periodontal therapy) were being observed by specialists. Logic might predict that this knowledge would have imposed a bias for the general dentists to perform more periodontal treatment rather than less. The other end of this spectrum would be that the specialists would "overtreat" and recommend more procedures than might be needed by the study patients now on maintenance recall with the general dentist. The study design does not allow for this difference to be resolved; only if patients were to be followed for a number of years could some information be obtained regarding true long-term treatment needs and outcomes.

The observation that the periodontists recommended similar amounts of surgery for both groups at the end of the study suggests that the use of the CHX chip did not result in a gain of periodontal attachment for the SRP + CHX chip group compared with the SRP group, a result that would be anticipated since the CHX chip is not indicated for periodontal attachment gain.¹² The observation further indicated that the periodontal evaluators did not detect any effect of the CHX chip as measured by their recommendations of treatment needs.

Differences in prosthodontic services. With respect to the differences observed in prosthodontic charges between the two study groups, the study data do not provide a clear explanation for the changes. One could speculate, however, that since the percentage of SRP group patients who underwent tooth extractions was numerically greater (nonsignificantly) than the percentage of SRP + CHX chip group patients who underwent extractions, SRP group patients may have received a greater number of prosthodontic services to replace missing teeth. This difference in charges for recommended prosthodontic services may have decreased at the 12-month evaluation at the academic centers because many of these missing teeth already may have been replaced during the year.

The periodontal evaluators did not detect any effect of the chlorhexidine chip as measured by their recommendations of treatment needs.

Comparison with past economic studies. Few economic studies of periodontal treatment have been published to date. A 1987 cost-effectiveness analysis hypothesized that nonsurgical treatments (including antimicrobial therapy) were more cost-effective than surgical treatments for patients with periodontal disease.¹³ Another study found that the use of counter-rotational brushes reduced the cost of periodontal treatments compared to a group using manual brushes.¹⁴ A Swedish study estimated dental care costs for periodontal and prosthodontic treatments in that country over a seven-year follow-up period; the largest proportion of those costs were due to prosthodontic treatment.¹⁵ A 1996 analysis evaluated the cost-effectiveness of new diagnostic tests for dental patients.¹⁶ This study found that diagnostic tests were not cost-effective for the general population, but for some subgroups (such as older patients, patients referred to periodontists and patients with a history of periodontal disease), diagnostic tests improved clinical outcomes.

Most recently, we reported a modeled economic analysis, which compared the same treatment alternatives as the current study (SRP alone and SRP + CHX chip) but which was based on data from a patient cohort followed for nine months, as well as clinical expert opinion of treatments that would be expected for those patients in months 10 through 12.⁶ The model predicted that the SRP + CHX chip group would receive significantly more maintenance treatment and significantly less periodontal surgery than the SRP group and that the average annual cost per patient would be essentially the same in the two treatment groups.

This study’s contributions. Our study makes a unique contribution to the dental literature in general, and specifically in regard to treatment of periodontitis. To our knowledge, this study is the first randomized clinical trial ever performed in dentistry that specifically examined patient treatment in general practice, capturing both the types of treatments performed and the costs of those treatments. Among its strengths were its size, randomized design and geographical distribution across the United States. Additionally, this study was conducted in general offices to reflect private dental practice as much as possible, but it also was overseen by blinded academic evaluators.

For general-practice patients with chronic periodontitis, our data indicate that more than one-half the costs of dental treatments for one year are for periodontal services. The adjunctive use of the CHX chip during that year had an effect on outcome in terms of treatments that were performed. We had suggested previously⁶ that for the CHX chip to help reduce periodontal surgery, nonsurgical treatment must be continued for a longer period (perhaps nine months) than in what may be the typical approach, in which patients might be re-evaluated for surgical needs six weeks to three months after undergoing SRP. Within the limitations of its design, our study has agreed with this hypothesis. Since general dentists refrained from recommending periodontal surgery or referring patients to periodontal specialists until they had undergone at least six months of study treatment, the reduction in the amount of surgical services performed was realized after at least six to nine months of non-surgical therapy, for patients treated both with and without the adjunctive use of the CHX chip.

Lindhe² also suggested that nonsurgical therapy may be continued successfully for prolonged intervals, ultimately resulting in a reduced need for surgery. Our data suggest that, along with the time interval, the addition of the CHX chip to the nonsurgical armamentarium may result in an even greater reduction in the probability of surgical procedures’ being performed for patients with periodontitis treated in general practice.

The CHX chip as a treatment option. It has been suggested that the CHX chip offers a new treatment option to achieve and maintain periodontal health.¹⁷ Since the CHX chip can be used easily by dental professionals, including both dentists and dental hygienists, and placed in less than a minute, it should be easily incorporated into both general and periodontal offices with minimal disruption of current patient scheduling practices.¹⁸ Our study has indicated that if the CHX chip is used routinely in general practice as an adjunct to SRP procedures, with treatment repeated every three months as indicated and with continuation of nonsurgical therapy for at least six to nine months, periodontal patients in general practice can anticipate about a 50 percent reduction in the likelihood of undergoing surgical procedures over the course of one year. This reduction is consistent with the approximately 50 percent incremental clinical benefit with respect to mean PPD reduction obtained after the use of the CHX chip as an adjunct to SRP compared with SRP alone in U.S. clinical trials.⁴

Implications for third-party payers. The primary perspective of this analysis was that of a comprehensive third-party dental payer providing full coverage. Since cost containment is being increasingly emphasized in this country, the results of this study have implications for third-party payers. The data suggest that for a period of one year, treatment costs for patients in general practice would be reduced compared with treatment costs in specialty practice. This suggestion results both from the increased fees for similar procedures performed by specialists vs. generalists and from the differences of the types or numbers of procedures recommended. The data do not allow for speculation regarding long-term costs. It is unclear whether unmet treatment needs, as suggested by the periodontists’ recommendations at the end of one year in this study, then would result in increased long-term costs because of further periodontal breakdown.

	CONCLUSIONS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
We conclude that over a period of one year in general practice, adjunctive use of the CHX chip resulted in a significant increase in total treatment charges. General practice patients who had chronic periodontitis and who were treated with SRP + CHX chip experienced a 50 percent reduction in the likelihood of undergoing surgical procedures compared with patients who were treated with SRP alone. The reduction in other dental treatment costs was sufficient to offset about one-half the acquisition cost of the chip. Periodontal specialists recommended that the majority of patients needed additional periodontal treatment, including similar amounts of surgery, at the end of the study year, regardless of the treatment group in which they had been placed.