Material and methods

This systematic review was conducted using best-practice principles and with guidance from previous reviews exploring research outcomes. A preliminary scope of the literature confirmed that no reviews were planned or published on this topic. Stakeholder opinion was sought from providers of hypodontia care and people with hypodontia in Leeds Dental Institute, Leeds, United Kingdom, during protocol development to maximize the relevance and applicability of this review. The protocol, published on PROSPERO in May 2015, was devised with a broader scope from which the identification of hypodontia care outcomes was 1 component.

The search strategy was deployed across 6 electronic databases using search terms related to hypodontia and treatment methods, plus 4 further electronic databases searched using terms for hypodontia only ( Table I ). No language restrictions were used. Additional grey literature searching was performed for unpublished or nonacademic literature. Hand searching of reference lists of relevant publications and reviews, citation searching for relevant papers identified in searches, and personal contacts via e-mail with the members of 2 dental specialty societies were also undertaken. Search terms were developed through piloting alternative search concepts based on population, intervention, and outcome to estimate sensitivity and specificity. Two concepts were used for searching based on MeSH terms and free headings for hypodontia and treatment methods ( Supplemental Table I ).

We developed inclusion and exclusion criteria for study selection ( Table II ). Studies including participants with syndromic hypodontia and cleft lip and palate were excluded due to potential differences in treatment options and care provision. Mixed-method study designs were included to ensure comprehensive identification of outcomes relating to care evaluation. Quality improvement methods such as audit and service evaluations, although distinct from research, were included as potential sources of information about stakeholder priorities and expected care standards. These records were managed separately during data extraction and synthesis.

Titles identified through the electronic database searches were collected and imported into EndNote X4 (Clarivate Analytics, Philadelphia, PA). Programmed deduplication was undertaken and checked manually using a sample of studies to ensure that the automated process was reliable. References were screened initially by title by a reviewer (S.B.) with verification of excluded titles by a second reviewer (B.K.). No errors were identified. Abstracts for all relevant titles were retrieved and evaluated independently by 2 reviewers (S.B. and B.K.). Full-text articles were obtained for studies meeting the inclusion criteria or when it was not possible to make a decision based on the abstract. Potential articles found through grey literature searches were included at this stage. Two reviewers (S.B. and B.K.) reviewed the full-text articles independently, and disagreements were resolved by consensus. Excluded studies were indexed in Excel (version14.7.4; Microsoft, Redmond, Wash) with the reason for exclusion.

A data extraction form modified from the Cochrane Public Health Group template was used to systematically extract data from each study. Coding was piloted for suitability before the review and developed through an iterative process until the form was judged fit for purpose. The following data items were extracted: author, date, design, setting and country, participants, interventions, comparator, outcome, measurement tools, reference for tool, and follow-up period. Data extraction was completed by 1 reviewer (S.B.) and checked for accuracy and consistency by a second reviewer (B.K.). Corrections were recorded to check for areas of ambiguity that would indicate internal inconsistency.

Synthesis of the studies was undertaken using a novel categorization approach based on the perspective of the evaluator in the study. Four outcome categories were developed: 3 subjective outcomes grouped by assessor (patient-reported, clinician-reported, and lay-reported) and 1 objective outcome (clinical indicators). Within each outcome category, themes and subthemes were used to separate the outcomes. For example, the patient-reported outcome category included themes for quality of life, appearance, function, and service delivery. In the appearance theme, the outcomes were further divided into subthemes including dental appearance and smile attractiveness. This coding theme was adopted to emphasize the use of patient-centered outcomes and to aid future work in developing outcomes of relevance to patients, clinicians, and stakeholders. This differs from the method previously reported in reviews of research outcomes where the outcome domains concerned outcomes from treatment in terms of efficacy.

In line with a previously published review exploring research outcomes, no synthesis of efficacy data was planned. Consequently, a methodologic quality appraisal of studies was not undertaken because the use of outcomes might not be associated with the quality of the studies.

Results

Electronic searches were initially completed on September 1, 2015, and were updated on December 19, 2016. The PRISMA flow diagram of study selection is shown in the Figure . Totals of 56 research records and 8 quality improvement records were included in the review ( Supplemental Tables II-IX ). The research reports included 1 randomized and 2 nonrandomized controlled trials and 53 observational studies ^{, , ,} ( Table III ). Research activity in hypodontia has grown with a noticeable increase in publications from 2000 onward. The majority of studies (43 of 56 [78%]) originated from Europe. The research teams had dental backgrounds and were based in a university or hospital with a few based in private practices. All studies were published in dental journals. Details of the key characteristics of the studies are summarized in the Supplemental Tables, grouped by intervention. Studies evaluated a single treatment method (n = 41) or compared methods (n = 12). Only 3 studies evaluated service delivery rather than outcomes from specific treatment methods ( Table IV ). The 8 quality-improvement reports were clinical audits undertaken in National Health Service hospitals and published between 2002 and 2016. The purpose of the audits was evaluation of service delivery including treatment duration, clinical efficiency, treatment planning, record keeping, and access to care ( Table V ).

Flow diagram of the study selection process.

Discussion

Patient-centered delivery of care and shared decision-making in treatment planning require provision of information that is useful and relevant to patients, yet this review illustrates inadequacies in the selection of appropriate outcome measures. People with hypodontia and their families are often committing to long courses of complex treatment and to enable effective care delivery, patients need to understand likely treatment consequences based on evidence with patient-centered outcomes. This review is unique in categorizing the outcome measures that are currently being used to evaluate treatment methods and service delivery in hypodontia. Understanding the characterization of potential outcomes and providing a system for categorization will assist future development of outcomes that satisfy all stakeholders in future studies.

The various outcomes identified from existing research predominantly assessed the efficacy of specific treatments, yet few studies indicated how the selection of outcomes and subsequent findings were expected to translate into improvements in clinical care. The lack of consistency in outcome measures limits the scope for comparing and combining study findings. Currently, the choice of outcomes appears to be largely driven by dental professionals and researchers. No study reported patient or public involvement in research design and selection of outcomes, and few teams included health-services research methodologists. This observation has 3 key implications: (1) the information gained from research may have limited value for patients during decision making; (2) there will be challenges when attempting to synthesize the evidence base for translation into practice; and (3) the research findings are unlikely to drive improvements in health services.

Clinical indicators were dominant throughout the studies, particularly for evaluating dental health after treatment and success of treatment. Clinical indicators can provide information about the effectiveness of treatment, the findings may have limited relevance for patients if a high level of dental knowledge is required for interpretation. Outcomes relating to quality of life and appearance are likely to be relevant to patients, because both have been shown to be motivators for seeking treatment and to be a determinant of satisfaction with treatment. It is less clear whether patients consider mastication and temporomandibular dysfunction useful measures of function and harm from treatment. A number of potential outcomes that could provide useful patient information for decision making were absent, such as patient preferences for dental health, treatment success, and further outcomes associated with service delivery. Clinician-reported outcomes were surprisingly scarce, limited to judgments on appearance after treatment. The beliefs, preferences, and unconscious biases of treatment providers could influence patient decision making, and evidence about clinician perceptions would be useful.

Perhaps the most surprising result was the almost complete absence of outcomes from a service-delivery perspective. Health-service research outcomes, such as measures of access to care, acceptability of treatment, and cost-effectiveness are essential for driving change. Quality-improvement studies were included in the review, with the expectation that the reports would indicate some outcomes that are important to service providers. Although the audits did provide some additional outcomes, these were limited to clinical indicators for measuring service delivery. Inadequate evidence for evaluating service delivery is a barrier to improving services and wider health. The prominence of evidence-based dentistry and the development clinical guidelines have resulted in emphasis on randomized controlled trials to measure the efficacy of treatments. This may in part explain the reduced focus on service evaluation and the scarcity of health-service research values in dental research. Present-day outcomes are unlikely to be able to answer important service delivery questions, such as whether a discrepancy exists between different providers of dental treatment, the effect of the dental care system on the health of hypodontia patients, and how social factors and organization structures may impact care provision. To enable recommendations to be made for improving hypodontia care, a broader approach to selecting research outcomes is required to incorporate information about the organization, management, finances, and delivery of care.

Trends were noted in the study characteristics and outcomes that reflect changes in research philosophy and the paradigm shift toward patient-centered care. Research activity has increased over time with a move toward assessing patient-reported outcomes. Oral health-related quality of life was first used in 2009 but has since been used in 4 further studies. The use of qualitative methods to gain a deeper understanding of patient experience was only recently reported but is likely to continue to gain popularity. Clinical indicators have remained popular over time, but there are signs of progress toward using better-defined, standardized measurement methods. Use of a standardized and validated tool is a variable across all outcomes and although some outcomes, such as oral health-related quality of life, tend to consistently use a widely accepted measurement tool (Oral Health Impact Profile), the tools for other outcomes, such as appearance, remain inconsistent.

In all studies, the construct of the research question and subsequent study design will influence the suitability and selection of outcome measures. Although randomized controlled trials are preferable for determining the effectiveness of a treatment, other research questions may demand different study designs : eg, surveys using qualitative methods to understand patient experience. In some instances, retrospective studies may be the only feasible design—eg, for determining long-term survival of deciduous teeth—but outcomes in these studies will be limited by the availability of the information. Regardless of study design, coherence is needed between the purpose of the research, the underlying theoretical basis, and the research design. The suitability of the outcome is intrinsically linked to the setting, participants, intervention, comparator, and methods. The location of studies, composition of the research team, and health care system in which the results will be applied are likely to influence the choice of outcomes. Universities may be more driven by academic interests than practice-based research teams, which will often be motivated by a need for clinical information or incentives for change.

This review is strengthened by the rigorous methodology and originality in subject and approach. The review was conducted following best-practice guidelines, and the findings will inform future research regarding development of patient-centered evidence for shared decision making.