Comparison of Splinting Interventions for Treating Mallet Finger Injuries

1. Literature Review: Mallet finger is defined as a loss of continuity of the distal insertion of the extensor tendon at the finger tip. It is a common hand injury in ball sports (McCue and Garroway 1985) but can also occur from minor incidents such as bed-making and trips/falls (Abouna and Brown, 1968). The injury results in a drooping of the distal inter-phalangeal joint, and is usually managed conservatively by splinting in hyper-extension for 6 or more weeks. This position allows relaxation of the tendon and encourages healing by bringing the torn ends or fracture fragments closer together during the healing phase. A recent Cochrane Systematic review of the evidence in treating this injury (Handoll and Vaghela, 2005) found that there is insufficient evidence in existing randomised controlled trials to establish the effectiveness of different (either custom-made or off-the-shelf) finger splints for treating mallet finger injury. They commented that there were only 4 trials that met the inclusion criteria, and all of these were "small, heterogeneous, inadequately described and reported….and had methodological flaws". Evidence provided is therefore an inadequate base for clinical decision making.
2. Rationale for project: If mallet injuries are not managed correctly, the patient can be left with a persistent extension lag (loss of voluntary straightening) and swan neck deformity (severe flexion deformity of distal finger joint plus a secondary hyperextension deformity of the proximal joint resulting from an imbalance in the extensor mechanism). Mallet injuries are relatively common, with an estimated 100 cases treated at the Alfred last year, and there is some debate about the best approach for their management. The plastic surgery unit is keen to determine the optimal treatment for these injuries. The Cochrane systematic review (Handoll and Vaghela, 2005) found that there were only four relevant trials that met inclusion criteria, and all of these had methodological flaws. As this is a commonly presenting injury at The Alfred, an opportunity exists to add to the evidence base by conducting a well-designed and thorough trial comparing the two most commonly advocated treatments with a suitable control.
3. Hypothesis/research questions: Null Hypothesis: There are no differences in outcome between different methods of conservative splinting treatment for mallet finger.
4. Aims: To determine the most appropriate conservative treatment for mallet finger injuries.
5. Methodology: We aim to conduct a randomised controlled trial of over 100 participants, based on current referrals to hand therapy. Subjects will be allocated to one of the following groups using a randomised computer sequence:

• Circumferential thermoplastic thimble splint, or
• dorsally applied aluminium-foam "Mexican hat" splint, or
• control splint (off-the-shelf thermoplastic "stack" splint with tape). NB it was considered unethical to provide a "no treatment" control given the amount of extension loss expected in an untreated injury and the likelihood of persistent finger deformity. All other aspects of treatment including frequency of contact, duration of initial period of splintage, and patient advice (eg hygiene routine, exercise, splint donning/doffing) will remain the same across all groups. Baseline data:
• age
• gender
• hand dominance
• time since injury
• type of injury (bony versus soft-tissue)
• degree of extensor lag (measured with a standardised goniometer)
• presence of other hand injuries on injured hand
• smoker/non-smoker
• medication
• hand dominance
• occupation or hobbies Progressive measurements (to be taken at 6, 8, 10, 12, and 20 weeks):
• degree of extensor lag (measured with a standardised goniometer)
• development of complications, and whether a change in splint type was required
• patient adherence to treatment protocol, based on self-report. Patients will be provided with a simple diary to complete which records incidences of removal or adjustment of splint, the reason(s) for this (eg skin breakdown, poor splint fit, dislodged during activity), and time out of splint Outcome measures (to be taken at 10, 12 and 20 weeks):
• degree of extensor lag at the DIP
• active range of motion at DIP
• patient satisfaction with result on 5-point likert scale
• pain, measured by 10 point Visual Analogue Scale Blinding of assessors:
• all progressive and outcome measurements will be taken by a separate assessor (not one of the current treating hand therapists) trained in taking the measurements who has been blinded to the type of splint used. To ensure this, splints will be removed prior to the assessment and re-applied afterwards by the Allied Health Assistant using a standard donning/doffing technique.

1. Inclusion/exclusion criteria: All referrals of patients with mallet finger injuries will be considered for inclusion with the exception of the following exclusion criteria:

• open injuries (where skin has been breached)
• mallet injury to thumb
• co-existing rheumatologic illness
• time from injury to presentation greater than 2 weeks

1. Randomisation procedures: Participants will be randomised to either of the 3 groups using a computer-generated randomised sequence.
2. Statistical or other analyses:

Sample size:

We are aiming for 100 subjects, but if referrals are low, we have calculated statistical power for a total of 75 subjects:

• With 25 subjects per group this study will have an 80% power to detect a difference in continuous variables equivalent to 80% of a standard deviation with a 2-sided p-value of 0.05. Based on the assumption of normality, a reduction of one standard deviation would be equivalent to about 24%, therefore a reduction of .8 of a standard deviation would be approximately equivalent to a 20% reduction. A difference of this size is perceived to be of clinical importance. Repeat measures analysis will further increase the power of this study although until between measures variability can be assessed, it is unsure by how much.

All data will be analysed using SAS version 8.2 (SAS Institute Cary, NC, USA). Data will be assessed for normality and log-transformed where appropriate. Univariate analysis will be conducted using chi-square test for equal proportion, analysis of variance and non-parametric Kruskal wallis tests where required. Multivariate analysis will be performed using generalised linear modelling adjusting for potential covariates and repeat measures. A two sided p-value of 0.05 will be considered to be statistically significant.