A prospective randomised comparative study of dynamic, static progressive and serial static proximal interphalangeal joint extension orthoses

Abstract

Introduction

Many different types of proximal interphalangeal (PIP) joint extension orthoses exist, yet evidence guiding orthosis choice is largely theoretical. The primary aim of this study was to evaluate the clinical effectiveness of three different PIP joint extension orthoses, over 4 weeks of treatment. Secondly, we aimed to explore the relationship between an abbreviated version of the Weeks test (WT) assessment of joint stiffness, and treatment outcome. Lastly, we wished to better understand participants’ satisfaction with orthotic treatment.

Methods

Using a randomised comparative study design, 61 participants were allocated to a serial static, dynamic or static progressive orthosis, 50 had follow-up data. Blinded assessment of function was completed before and after 4 weeks of orthotic intervention and a standardised therapy program. Participants were blinded to alternative groups.

Results

Baseline active PIP extension ranged from 14° to 65°. The average improvement in active PIP extension was −9.1° (95% CI −11.0°, −7.1°). There were no statistically significant differences in outcome between the three orthoses groups. However, a trend was observed with greater improvement in active extension for those in the dynamic Capener (−11.5°) compared to the static progressive belly gutter (−7.3°) or serial cast (−8.7°) groups, with less total end range time required. The abbreviated WT was significantly associated with improvement in active extension (p = 0.001). Participants reported a high degree of satisfaction with their orthosis regardless of type.

Conclusions

No single orthosis demonstrated statistically greater effectiveness, although the dynamic Capener orthosis appeared more efficient. The abbreviated WT is associated with treatment outcome.

Introduction

The proximal interphalangeal (PIP) joint is one of the most frequently injured joints in the hand.^1–3 Loss of extension or a fixed flexion deformity (FFD) is a common complication of PIP joint injury.^4,5 Evidence supports the use of mobilisation orthoses for the management of persistent PIP joint FFD.^6,7 Consequently, for many clinicians, orthoses have become the modality of choice for non-operative treatment. ⁸

Mobilisation orthoses apply gentle tension to the contracted joint at the end of the available range, promoting soft tissue growth and increased motion.^8,9 Total End Range Time (TERT) is the term used to describe the length of time an orthosis holds a joint under tension at the end range. Research indicates that greater TERT facilitates greater motion.^10,11 Three key classes of mobilisation orthoses are described in hand therapy literature; dynamic, static progressive and serial static (Figure 1). ⁸ Despite a common end goal, each type of orthosis has a different biomechanical action. Serial static orthoses are regularly remoulded by the therapist over time as the soft tissues in the finger gradually adapt. ⁸ Dynamic orthoses use an elastic mobilising force that is transmitted from a stable plastic base applying ongoing tension to the PIP joint whilst in situ. ⁸ In contrast, static progressive orthoses use an inelastic mobilising force that requires the patient to regularly adjust the tension, as the PIP joint gradually yields throughout the application. ¹²

Figure 1.

Left to right (a) serial static orthosis (cast), (b) dynamic Capener orthosis and (c) static progressive belly gutter orthosis.

The evidence available to assist therapists with clinical reasoning in choice of PIP joint extension orthosis is based on expert opinion and theory (level 5 evidence). ¹³ It has been theorised that there is a hierarchy of orthosis efficacy based on the assessment of joint stiffness, with static progressive orthoses considered superior for particularly stiff joints, followed by dynamic and lastly serial static orthoses.^12,14 The need for empirical research trials that evaluate this theoretical hierarchy and relevance to everyday clinical practice has been highlighted.^14–16 For example, an orthosis may perform well under the controlled circumstances of an experiment (efficacy) but not necessarily translate well into the broader clinical environment (effectiveness) due to contextual, patient-specific and therapist factors. ¹⁷

To explore those factors specifically influencing clinical reasoning and choice of PIP joint extension orthosis, Glasgow and Peters surveyed members of the Australian Hand Therapy Association. It was identified that therapists commonly used a wide variety of orthoses; however, there was no clear preference for any single type. Orthosis choice was primarily influenced by patient needs and therapists' experience and confidence with fabrication. ¹⁶

Benedi published a randomised study of prefabricated dynamic (Capener) and static progressive (safety pin) orthoses for PIP extension deficit. ¹⁸ They found no statistically significant difference between groups; however, due to the small sample size (n = 17), it was not possible to exclude failure to identify a significant difference due to low study power.

Young et al. completed a systematic review in an attempt to answer whether static or dynamic orthoses were more effective for PIP joint extension deficit. These authors found that studies used heterogeneous methodologies, and therefore, they were unable to conduct a meta-analysis or pool results to be able to answer their question. They concluded that further empirical research was required to support clinical reasoning. ¹⁹

Assessment of joint stiffness using the Weeks test (WT) method has been found to predict response to orthotic intervention. ²⁰ The WT has good test re-test reliability and may assist decision making regarding the need for orthotic treatment or surgical intervention. ²⁰ The change in motion after 30 min of heat and stretch is the estimate of stiffness (WT score).^14,20 However, in a busy clinical environment, this test is time intensive and as such not always practical. ²¹ An abbreviated version of the WT would be useful if it could be demonstrated that it was similarly related to the likely treatment outcome.

Effective orthosis choice for PIP FFD remains a clinical conundrum.^16,21 The available theoretical evidence does not appear to be integrated into daily practice questioning clinical relevance. ¹⁶ The primary aim of this study was to evaluate the effectiveness of three PIP joint extension orthoses in everyday practice. Active PIP joint extension was chosen as the primary outcome measure. The secondary aim was to examine the relationship between an abbreviated version of the WT assessment of joint stiffness and improvement in active extension. Lastly, we wished to better understand participants’ satisfaction with orthotic treatment.

Methods

Trial design

We conducted a randomised comparative study. Using the PRECIS-2 (Pragmatic-Explanatory Continuum Indicator Summary) tool, this study sits on the pragmatic side of the continuum and was designed to represent orthosis effectiveness in everyday clinical practice (average score 4.1). ²¹

Participants

Participants with PIP joint extension deficits attending the urban public hand clinics at Queen Elizabeth II Jubilee Hospital, Princess Alexandra Hospital and the Royal Brisbane and Women’s Hospital in Brisbane, Queensland, Australia, between July 2016 and September 2020 who met the selection criteria were invited to participate in the study. There were no financial incentives for participation. Selection criteria included: passive PIP joint extension deficit ≥10°, ≥18 years of age, injured structures were stable and ready to commence mobilisation orthosis and able to provide voluntary written informed consent. Patients with central or peripheral nervous system dysfunction, acute complex regional pain syndrome, arthritis and/or infection were excluded.

Interventions: Orthoses

Three different orthoses were used, one from each mobilisation class, namely, serial static, dynamic and static progressive. These corresponded to those most commonly used by Australian hand therapists, namely, serial casts (static), belly gutter (static progressive) and handmade Capener (dynamic) orthoses ¹⁶ (Figure 1).

Outcomes

Active PIP extension was the primary outcome measure used. Passive extension, arc of motion and TERT were also studied. Patient-reported outcomes were examined using the Patient Rated Wrist and Hand Evaluation (PRWHE), ²² the Patient Evaluation Measure (PEM) ²³ and a customised orthosis satisfaction tool.

Data were collected on specific variables previously demonstrated to influence the outcome with orthotic intervention, namely, WT score, diagnosis, time since injury and TERT.^11,20 Additionally, demographic data were collected on age, gender, work status, dominance, hand injured, digit and whether surgery had occurred.

Sample size

The sample size was calculated using a two-sample t-test assuming equal variance for detection of 5° difference in active extension between at least two of the groups. In a previous study, we found an average improvement in active PIP extension of 2.88° after 4 weeks of extension splinting and an improvement in passive extension of 7.52°. ²⁴ Hence, an effect size of 5° was considered reasonable for this study. Assuming a common standard deviation of 6.062,^24,25 to reach a power of 80% and alpha of 5%, a total sample size of 75 was required with 25 participants in each group.

Randomisation and blinding

Before commencing study recruitment, a computerised random sequence was generated and stored with a therapy assistant not involved in data collection. Following written informed consent and completion of the initial participant assessment, the treating therapist contacted the assistant to receive orthosis allocation. In this way, the therapist was blinded to group allocation during the initial assessment. Likewise, the final assessment at 4 weeks was completed by an independent therapist who was also blinded to the treatment group. Participants were blinded to alternative groups.

Procedures

All therapists collecting data were trained by the primary investigator in the study procedures, including the construction of the orthoses as required. The initial assessment included a verbal history and demographic data collected from patient report, hospital medical notes and a review of imaging results. Cold active and passive range of motion was measured using a standard metal finger goniometer. The abbreviated WT was performed with participants applying a passive PIP extension stretch using their good hand to the affected proximal phalanx (Figure 2), avoiding hyperextension of the distal interphalangeal joint or undue pressure over the PIP itself. Both hands were then placed in a moist hot pack between protective layers of a towel for 10 min while the stretch was applied. Active and passive extension was re-assessed after 10 min and the change in each formed the abbreviated WT measures. On completion of the physical assessment, participants completed the PRWHE and the PEM. Allocation to the orthosis group was then obtained from the therapy assistant and the orthosis was fabricated by the therapist.

Figure 2.

Technique for proximal interphalangeal joint extension stretch for abbreviated Weeks test (performed inside heat pack).

Participants were instructed to wear their orthosis for 12 h a day regardless of design and all received an exercise program and oedema management with compressive wrap. Casts were designed to be removable. A paper diary was provided for participants to record compliance.^20,25 Participants were instructed to record honestly exercise frequency and orthosis use in their diary. Participants in the casting group were reviewed each week and their cast changed. Participants in the belly gutter and Capener groups were reviewed every 1–2 weeks. This frequency of treatment was consistent with clinical practice across data collection sites.

The final evaluation was completed after 4 weeks of treatment by a blinded assessor using the same procedure as the initial evaluation, with the addition of the participant satisfaction questionnaire. A study timeframe of 4 weeks was chosen to ensure that participants would have the opportunity to trial an alternative orthosis within a reasonable timeframe if needed. Previous research has identified that time since injury is a significant factor influencing outcome with orthotic intervention with slower progress expected as time increases. ²⁰ As little is known about the advantage/disadvantages of one orthosis design over another, we did not want to inadvertently compromise participants' long-term outcomes. After this final assessment, data collection for the study was complete and participants continued treatment as required.

Statistical analysis

Data were analysed in SPSS version 27 (IBM Corp, Armonk, NY). The comparison of interest was the effect of orthosis type on selected outcome measures. Categorical data were reported as frequency and percentage and continuous data were checked for normality and reported as mean and standard deviation. Categorical data were examined using the Pearson chi-squared test or Fisher’s exact test if more than 20% of the expected values were less than 5. Continuous data were analysed using a one-way ANOVA. Paired t-tests were used to examine the overall change for each of the outcome measures. An analysis of covariance was used to examine the change between week four and baseline for all outcome measures with baseline measures and orthosis group included as covariates. Univariate analysis for each outcome with participant characteristics was explored and variables identified as having p-values of 0.25 or lower were examined in multivariable linear regression using backwards selection. Participant characteristics that differed between orthosis groups were examined with modelling as well as factors known to influence treatment outcome. ²⁰ The purpose of the multivariable linear regressions was exploratory, to see if their inclusion in models changed estimates for the orthosis groups. Beta estimates and marginal mean estimates and their 95% confidence intervals are reported.

Results

Sixty-five patients were invited to participate in the study. There were no adverse events. Two patients declined to be involved and two were excluded, one patient with severe arthritis and the other due to peripheral neuropathy (Figure 3). Table 1 presents sample characteristics and baseline measures according to the orthosis allocation group. There were some statistically significant differences observed between groups at baseline (Table 1). Participants in the belly gutter group were on average 15 years younger than those in the handmade Capener group. They were also more likely to have injured their little finger compared to the other groups and had better active (26°) and passive (20°) baseline extension. The serial cast group averaged greater final TERT than the other groups despite all participants being told to use their orthosis for 12 h per day (Table 2). These differences observed between groups can occur in any randomised trial as a result of chance alone. Subsequently, these variables were included in the multivariate models used to assess for significance in relation to orthosis group and change in active extension. There was no statistically significant difference in joint stiffness between groups. The average abbreviated active WT was 6° indicating a high degree of stiffness for our sample. ¹⁴

Figure 3.

CONSORT diagram outlining research flow from study enrolment to data analysis.

Table 1.

Sample characteristics and baseline measures according to the orthosis allocation group.

Variables	Overall	Serial cast	Belly gutter	Handmade Capener	p-value
	n = 50	n = 18	n = 17	n = 15
	n (%)	n (%)	n (%)	n (%)
Age (years, mean (SD))	38 (13)	40 (13) a	30 (8)	45 (14) a	0.007
Gender					0.75
Male	33 (66.0%)	13 (72.2%)	11 (64.7%)	9 (60.0%)
Female	17 (34.0%)	5 (27.8%)	6 (35.3%)	6 (40.0%)
Employment					0.28
Employed	42 (84.0%)	13 (72.2%)	15 (88.2%)	14 (93.3%)
Retired/non-working/student	8 (16.0%)	5 (27.8%)	2 (11.8%)	1 (6.7%)
Hand injured					0.53
Left	26 (52.0%)	10 (55.6%)	7 (41.2%)	9 (60.0%)
Right	24 (48.0%)	8 (44.4%)	10 (58.8%)	6 (40.0%)
Dominance (n = 49)					1.00
Left	8 (16.3%)	3 (16.7%)	3 (18.8%)	2 (13.3%)
Right	41 (83.7%)	15 (83.3%)	13 (81.3%)	13 (86.7%)
Digit injured					0.007
Index/middle	12 (24.0%)	6 (33.3%)	1 (5.9%)	5 (33.3%)
Ring	14 (28.0%)	5 (27.8%)	2 (11.8%)	7 (46.7%)
Little finger	24 (48.0%)	7 (38.9%)	14 (82.4%)	3 (20.0%)
Time since injury (n = 49)					0.33
12 weeks or under	26 (53.1%)	8 (44.4%)	11 (68.8%)	7 (46.7%)
13 weeks or more	23 (46.9%)	10 (55.6%)	5 (31.3%)	8 (53.3%)
Diagnosis					0.73
Volar plate injury +/− #	23 (46.0%)	8 (44.4%)	9 (52.9%)	6 (40.0%)
Tendon/soft tissue	11 (22.0%)	4 (22.2%)	2 (11.8%)	5 (33.3%)
Other complex #	16 (32.0%)	6 (33.3%)	6 (35.3%)	4 (26.7%)
Surgery					0.034
No	23 (46.0%)	7 (38.9%)	12 (70.6%)	4 (26.7%)
Yes	27 (54.0%)	11 (61.1%)	5 (29.4%)	11 (73.3%)
Mechanism of injury					0.43
Dislocation (hyperext/flex)	26 (52.0%)	9 (50.0%)	10 (58.8%)	7 (46.7%)
Die punch/torsional injury	10 (20.0%)	3 (16.7%)	5 (29.4%)	2 (13.3%)
Soft tissue/other	14 (28.0%)	6 (33.3%)	2 (11.8%)	6 (40.0%)
Baseline measures (mean (SD) (min, max))
Active PIP extension (°)	32 (11) (14, 65)	34 (10) a	26 (8)	35 (13) a	0.04
Passive PIP extension (°)	26 (11) (5, 60)	28 (10) a	20 (10)	29 (13) a	0.045
Active arc of motion (°)	50 (16) (11, 80)	49 (15)	56 (16)	43 (18)	0.087
Passive arc of motion (°)	62 (17) (16, 97)	62 (14)	69 (17)	55 (17)	0.08
Baseline PEM total score b	50 (13) (24, 89)	51 (15)	48 (10)	50 (16)	0.83
Baseline PRWHE total score	29 (16) (2, 79)	33 (18)	30 (12)	25 (15)	0.37

Note: PEM: Patient Evaluation Measure; PRWHE: Patient Rated Wrist and Hand Evaluation; PIP: proximal interphalangeal.

min is minimum; max is maximum; (°) degrees; SD standard deviation.

^aMeans with the same letter are not significantly different within rows; # = fracture.

^bn=49 for baseline PEM score.

Table 2.

Improvement in outcome measures after 4 weeks of treatment according to the orthosis group.

Outcome measure	n	Overall change	Serial cast	Belly gutter	Handmade Capener	p-value∼
		Unadjusted mean (95% CI)^	Estimated marginal mean (95% CI)∼	Estimated marginal mean (95% CI)∼	Estimated marginal mean (95% CI)∼
Change in active extension (°)	50	−9.1 (−11.0, −7.1)	−8.7 (−12.1, −5.3)	−7.3 (−10.9, −3.7)	−11.5 (−15.2, −7.8)	0.28
Change in passive extension (°)	49	−8.8 (−10.7, −6.9)	−9.2 (−12.5, −6.0)	−7.5 (−11.0, −4.0)	−9.8 (−13.5, −6.0)	0.65
Change in active arc of motion (°)	50	12.1 (8.9, 15.4)	11.0 (5.5, 16.6)	12.2 (6.3, 18.1)	13.3 (7.1, 19.6)	0.86
Change in PEM	47	−8.5 (−11.7, −5.3)	−8.6 (−13.3, −3.8)	−7.0 (−11.9, −2.2)	−10.0 (−15.3, −4.8)	0.70
Change in PRWHE	49	−9.0 (−12.8, −5.2)	−8.1 (−14.0, −2.1)	−11.4 (−17.5, −5.3)	−7.2 (−14.0, −0.4)	0.60
Total hours orthosis used over 4 weeks (mean (SD) (min, max)	47	290 (96) (63, 544)	343 (107)	257 (58)a	262 (91)a	0.012

Note: PEM: Patient Evaluation Measure; PRWHE: Patient Rated Wrist and Hand Evaluation.

^ Paired t-test, p-values for all outcomes are <0.001; ∼ ANCOVA, min is minimum; max is maximum.

^a EStimated means with the same letter are not significantly different.

Table 2 shows the change in outcome measures according to group after 4 weeks of treatment. There were no statistically significant differences observed on outcome measures between orthosis groups for both the adjusted and unadjusted analyses. Although not statistically significant, participants in the handmade Capener group saw slightly more improvement on average for all outcomes examined except the PRWHE. Participants in the serial cast group had on average slightly more improvement in ROM than those in the belly gutter group, but the belly gutter group had slightly better arc motion than the serial cast group as well as a better score change on the PRWHE (Table 2).

Multivariable linear regression models were used to study the effect of the orthosis group on the outcome measures after adjusting for important independent variables. The number of exercise sessions performed was not included in the models due to significant missing data (not reported in Table 1). For the primary outcome measure, change in active extension, baseline active extension, orthosis group, gender, employment, active WT, time since injury and diagnosis went into the initial model. The final model contained baseline active extension, orthosis group, gender, Weeks active test and diagnosis, as shown in Table 3.

Table 3.

Comparison of unadjusted and adjusted models for the primary outcome measure – change in active extension.

Model parameters	Unadjusted			Adjusted
	Beta (95% CI)	Estimated marginal mean (95% CI)	p-value	Beta (95% CI)	Estimated marginal mean (95% CI)	p-value
Change in active extension (°)	n = 50			n = 50
Orthosis			0.28			0.23
Handmade Capener	−2.8 (−7.7, 2.2)	−11.5 (−15.1, −7.8)		−1.2 (−5.1, 2.9)	−11.2 (−14.3, −8.1)
Belly gutter	1.4 (−3.6, 6.4)	−7.3 (−10.9, −3.7)		2.6 (−1.6, 6.7)	−7.4 (−10.6, −4.3)
Serial cast	Reference	−8.7 (−12.1, −5.3)		Reference	−10.0 (−12.9, −7.1)
Baseline active PIP extension^	0.1 (−0.1, 0.3)		0.54	−0.0 (−0.2, 0.1)		0.55
Intercept	−10.6 (−18.1, −3.2)		0.001	3.1 (−10.9, 4.8)		0.55
Diagnosis						0.005
Other complex #	n/a			6.5 (2.7, 10.4)	−5.8 (−8.9, −2.8)b
Tendon/soft tissue	n/a			1.9 (−2.6, 6.4)	−10.5 (−14.2, −6.7)ab
Volar plate injury +/−	n/a			Reference	−12.3 (−14.9, −9.8)a
Active WT^				−1.0 (−1.6, −0.4)		0.001
Gender						0.014
Female	n/a			−4.5 (−8.1, −1.0)	−11.8 (−14.7, −8.9)
Male	n/a			Reference	−7.3 (−9.4, −5.1)

Note: PIP: proximal interphalangeal; WT: Weeks test.

^{a and b} estimated means with the same letter are not significantly different; ^ active WT centred at 6 and baseline active extension centred at 32.

The active WT was significantly associated with a change in active PIP extension (p = 0.001). Females improved on average −11.8° compared to males −7.3° (p = 0.014). Participants with volar plate injuries made the greatest progress over the 4 weeks (average −12.3°) followed by tendon and soft tissue injuries (−10.5°) and lastly participants with complex fractures (−5.8°) (p = 0.005). The inclusion of these other factors in analyses did not significantly change marginal mean estimates for the orthosis groups. The models for the change in passive extension, arc of motion, PEM and PRWHE were not significant. The total pooled group of participants demonstrated a statistically significant improvement on all outcome measures when orthosis type was not a consideration (p < 0.001).

There was a high degree of satisfaction with each orthosis and therapy provided with no statistically significant differences observed between orthosis groups for all questions except for whether the ‘orthosis worked well’ (Appendix 1 online supplementary file). Seventy-two per cent of participants wearing the serial cast were satisfied that it straightened their finger compared to 100% wearing the belly gutter and 86.7% wearing the handmade Capener. Participants in the serial cast group believed that they were more likely to be able to sleep in their orthosis (94.4%) than those in the other groups although this difference was not statistically significant.

Discussion

This study aimed to compare the effectiveness of three commonly used extension orthoses for PIP joint FFD in everyday clinical practice. A secondary aim was to explore the relationship between an abbreviated version of the WT and response to therapy, as the first step towards examining clinically utility. Participants’ satisfaction with orthotic treatment was also examined.

No statistically significant differences between the three orthoses groups used in this study were observed. Consequently, our findings do not support the notion that any single orthosis is superior. They also contradict participants’ perceptions that the belly gutter orthosis ‘worked well to straighten my finger’ (100%). Objectively, the dynamic Capener orthosis appeared to perform better followed by the serial cast. Yet participants had lower perceptions of the effectiveness of these two orthoses (Capener 86.7%, serial cast 72.2%) compared to the belly gutter. The belly gutter orthosis used in this study was able to be progressively tightened by participants consistent with static progressive designs using a Velcro strap (Figure 1). It is possible that greater stretch was used by participants in tightening this orthosis to apply the mobilising force resulting in the impression that the orthosis was effective. Too much force however results in micro-tears, inflammation and further fibrosis, perpetuating the cycle of joint contracture. ²⁶

Participants demonstrated a positive response to treatment overall when the orthosis group was not a consideration. How much of this improvement was due to splinting, exercises or time is difficult to determine. Research directly comparing the effectiveness of an orthotic intervention to other standard treatments (e.g. exercises and functional activities) for the management of PIP joint FFD is not available for comparison.

The active version of the abbreviated WT was significantly associated with improvement in active PIP joint extension as previously found with the longer 30 min test. ²⁰ The greater the change in active ROM after 10 min of heat and stretch with the abbreviated WT, the better active extension would be after 4 weeks of treatment. Consequently, the abbreviated WT may provide a more suitable means of examining the likely response to treatment in the busy clinical environment. A larger abbreviated WT score would suggest a trial of therapy is indicated. A smaller WT score would suggest a less compliant joint that may require surgical intervention as opposed to engaging in a lengthy period of extensive therapy. Test re-test reliability of the abbreviated WT needs to be the subject of further research to determine if is comparable to the 30 min WT.

Females made significantly greater gains in PIP extension than males. The reason for this was unclear with subsequent analysis showing no difference between men and women in terms of joint stiffness, TERT, age or affected digit. One possibility would be that men generally have larger joints that may require greater force to promote tissue growth, compared to women. Force was not able to be controlled for in this study due to the differences in orthosis biomechanics. For example, static progressive orthoses are designed so that participants can self-adjust the applied force as the joint gives, in contrast to serial casts which are unable to be adjusted at all once in situ. Further research examining optimal force parameters for mobilisation orthoses is needed.

The poorer progress of participants in the complex fracture group was not surprising due to involvement of the articular surface of the PIP joint and the bony interface. Participants in the other groups were more likely to experience restriction in motion resulting from soft tissue limitations, which are more amenable to therapy. This finding is consistent with that of previous authors. ²⁰ Participants in the belly gutter group were on average 15 years younger and more likely to have injured their little finger with better baseline extension. Age, digit and baseline active extension were not related to final active extension in our sample.

Participants in the serial cast group were more able to sleep in their orthosis (94.4%) than those in the other two groups (Appendix 1 online supplementary file). This corresponded to a significant difference observed between groups in TERT, with those in the cast group wearing their orthosis for 343 h compared to 257 in the belly gutter and 262 in the Capener group. As TERT is positively associated with improvements in ROM, this may contribute substantially towards the effectiveness of serial casting.^10,11 Casts are low profile and unobtrusive and do not hinder motion at the MCP joint which makes them easy to wear during everyday activities. They are easy to fabricate, adding to their popularity in clinical use. ¹⁶ However, in terms of efficiency per number of hours the orthosis was worn, the dynamic Capener appeared to perform best in our sample, with less TERT resulting in the apparent best final active extension. Perseverance in achieving proficiency in Capener fabrication appears justified.

Limitations

The most significant limitation of this study was that the projected sample size was not reached reducing statistical power. The largest difference between groups for a change in active PIP ROM after 4 weeks of treatment was 4.2^o with a common standard deviation of 10.4. This is below the difference we would like to see in practice (5^o) and we did see more sample variability than we expected. An increased sample size is likely to have decreased variability; however, it may still not have resulted in a statistically significant difference between groups. However, statistical significance does not necessarily reflect clinical importance or effectiveness in everyday practice.^17,21

This study received little external funding and without a dedicated research assistant it relied on the good will of busy clinicians and colleagues to drive data collection. It was ceased in 2020 during the Covid 19 pandemic as participants were not able to attend follow-up appointments. Amongst other government mandates health services were instructed that clinical research be suspended to reduce foot traffic in/out of hospitals and to divert resources to inpatient care. These factors contributed to the final sample size.

Another limitation of this study was that participant progress was evaluated over a short period of time, 4 weeks. This was a conscious decision to provide participants with an opportunity to trial an alternative orthosis within a reasonable timeframe if their allocated orthosis proved unsuitable, as little was known regarding the merits of one design over another. Future research with a longer follow-up timeframe would be useful to build on our preliminary findings.

The use of a written self-report tool to collect data on TERT and exercise frequency was another limitation due to missing data. Again, future research should consider alternative methods of recording compliance (e.g. digital apps/electronic platforms).

Conclusions

No orthosis demonstrated superior effectiveness in our sample of stiff PIP joints although the Capener orthosis appeared slightly more efficient. The abbreviated active WT appears to be associated with final active extension; however, future research should examine reliability. Therapists should consider their skill in orthosis fabrication and match this with patient-specific needs to justify orthosis choice. Factors such as adequate TERT, comfort, biomechanical effectiveness and compliance with treatment are paramount to ensuring a positive outcome. Further research is recommended, to pool findings from this study with those of other authors.

ORCID iD

Celeste Glasgow https://orcid.org/0000-0002-3356-4983