Study design and setting
The ConVenTu study is a prospective multicenter, randomized study conducted in clinical settings in seven hospitals located in all four Regional Health Authorities in Norway. The postoperative care is committed by either an otolaryngologist at the hospital where they had VT surgery, or by the patients’ regular GP. A flowchart of design and timeline in the study is presented in figure 1, and a SPIRIT schedule of enrolment, interventions, and assessments in figure 2.
The Department of Otolaryngology, Head and Neck Surgery at St. Olavs Hospital is the main investigator site and coordinator of the study. The study is performed in collaboration with the Norwegian University of Science and Technology (NTNU). The project team leader and main project coordinator have regularly communication. Study visits have been performed at all study sites. In addition, meetings with all study sites are held annually or more frequently.
Eligibility criteria
Eligible to participate are children, age 3-10 years, where VTs are placed in one or two ears. Exclusion criteria are conditions that are in need of closer postoperative care by otolaryngologists; for instance medical syndromes (e.g Downs syndrome), other coexisting severe diseases, or severe neurogenic hearing loss in at least one ear (we have excluded children with > 50 dB hearing thresholds in at least one frequency 0.25 - 4.0 KHz). Also, children with cognitive impairment or no comprehension of Norwegian language are excluded.
Random allocation of postoperative care
At discharge from hospital, each patient that agreed to participate, is randomly allocated to receive postoperative care either by an otolaryngologist or by their regular GP (Figure 1). The randomization procedure is computerized and carried out by means of the ‘WebCRF’ (Clinical Registration Form) software developed at Unit for Applied Clinical Research at The Faculty of Medicine and Health Sciences, NTNU (15). The design is block randomization with varying size of blocks, with an allocation ratio of 1:1, stratified on study center.
Procedures for postoperative controls
Children randomized to postoperative care by otolaryngologists will get an appointment at 6, 12 and 18 months after surgery, to reflect the existing management of these patients in Norway. Postoperative care by GPs will not be at fixed appointments, but on the initiative from the guardians as this reflects existing management in general practice. Guardians of children randomized to the GP group will receive written information on when to contact their GP; for instance, if the child has otorrhea, persistent reduced hearing, or otitis media. The GPs receive information about the study and a procedure for postoperative care and how to treat the most common complications. In addition, a procedure for handling of complications is enclosed in the discharge report after surgery, so it is available for the GP at point-of-care. If needed, as judged by the GP, the children will be referred to an otolaryngologist. If the VTs have not been spontaneously rejected within 18 months, the GP is requested to refer the patient to an otolaryngologist. Accordingly, the guardians are told to contact the GP latest 18 months after surgery (Figure 1).
Assessment of clinical and sociodemographic factors
All participants will be evaluated with audiological tests and questionnaires before surgery and two years after. Audiological tests are performed with audiometry and tympanometry. Hearing thresholds are measured by pure tone audiometry. In the youngest children equipment for play audiometry is used if needed (16). Results from at least three of the pure tone thresholds in dB at 0.5-1-2-4 kHz form the pure tone average (PTA). Two years after surgery all participants will be examined by an otolaryngologist. Complications during the postoperative period will be registered by carefully examining the patient record.
Sociodemographic information is assessed at enrollment. At the end of the follow-up period the guardians evaluate the postoperative care. Both before and two years after surgery the guardians are asked to complete questionnaires regarding their perception of the child’s HRQoL (see secondary outcome).
Each study center has a local coordinator responsible for a Case Report Form to ensure complete recording of individual data.
Outcomes
PTA in dB is defined as primary outcome variable.
The following clinical and sociodemographic factors are considered as secondary outcomes:
- Middle ear function assessed with tympanometry and otomicroscopy (17)
- Complication rates
- Disease specific HRQoL by the otitis media questionnaire for children (‘OM8-30’) (18)
- Generic HRQoL by ‘PedsQL’, based on proxy-report (participants aged 3-8 years) or by adolescent self-report (8-12 years) and ‘SDQ-Nor’ (19, 20)
- Guardians evaluation of the postoperative care
The secondary outcome measures no. 1-2 act as surrogate markers for possible reduced hearing in the future. Complications after insertion of VTs include otorrhea, occlusion of tubes, premature extrusion, persistent perforation in the tympanic membrane, retraction pocket and cholesteatoma (6, 7, 21). The secondary outcome measures no. 3-5 is assessing the risks and benefits of postoperative care and important for patient-centered care.
Participant timeline
See Figure 1 and Figure 2
Sample size and power
We defined equality in the main outcome variable (PTA two years after surgery) as < 5dB, and 5dB as the minimal clinically important difference (equivalence margin), since measurements of hearing thresholds are determined in 5dB steps (22). To avoid an incorrect conclusion about no difference between the groups (Type II error), the power of the tests for detecting a difference of ≥ 5 dB must be high (≥ 95 %, error margin <5 %). Thus, with a power of 95% and a significance level of 5% (two-sided test), 105 participants is needed in each group to be able to detect an absolute difference in PTA between the groups of ≥5 dB (standard deviation of 10 dB in each group). With additional 15% for dealing with potential skewly distributed variables, and further a 20% increase to account for potential dropouts during the follow-up period, we are left with a sample size of 145 participants in each group. To maintain power in an analysis stratified for seven study centers (six additional parameters in model), the total sample size must be increased by at least 60 participants, applying general rules of thumb of at least 10 cases per parameter. This leave us with a sample size of at least 175 participants in each group if we assume that the 60 participants are evenly distributed in the two randomization groups. We chose to truncate upwards to 200 participants in each group (Figure 1) to maintain power in case of a higher dropout rate or a need for additional adjustment factors in the model.
Statistical analysis
The analyses will be performed according to intention-to-treat. A general linear model and/or a linear mixed model (LMM) will be applied to compare PTA and mean change in PTA after two years. Similar methods will be applied for other relevant variables that are measured on a continuous scale (HRQoL). Log-transformation, or non-parametric methods, may need to be considered. Generalized linear mixed model, and/or McNemar’s test is relevant for analyzing changes in categorical, dichotomous outcome variables. A Chi-square test is relevant for comparing the two study groups with respect to number of complications during postoperative care. Equality between the two treatment groups will be evaluated in terms of the magnitude of observed differences (point and interval estimate).