CHiropractic for Infantile Colic Study (CHIC)

Overall Design

The study will be a two-by-two factorial design combining groups of chiropractic treatment or 'Best Standard Care' with blinded or unblinded parents.

The two treatment groups will be 'Best Standard Care only' (BSC) and 'Best Standard Care plus chiropractic' (Chiro). Best standard care will be based on the recommendations from United Kingdom (UK) National Health Service (NHS), using the Map of Medicine care map. Chiropractic treatment will be delivered according to the protocol defined by the participating chiropractors.

There will be two blinded and two unblinded groups. The blinded BSC group will receive the 'sham' treatment in addition to the BSC counselling. The unblinded BSC group will not require this.

In addition, to allow the effects of the n-shaped curve inherent in the natural course of colic to be evaluated, all groups will be stratified by age at entry (below 4 weeks (expected increasing crying over the 2 weeks of the main intervention), between 4 and 6 weeks (expected plateau), and over 6 weeks (decreasing crying)).

The study has a longitudinal design over a ~4 1/2 week period.

For the primary outcomes which take data from the crying diaries, the main data points are defined as:

• Baseline: the 24 hour period of the day before the first treatment
• First treatment: the 24 hour period of the day after the first treatment
• Mid-point treatment: the 24 hour period of the day after the second treatment
• Final treatment: the 24 hour period of the day after the third treatment

Secondary outcomes will include data from the following points:

• Follow-up 1: the 24 hour period of the day 7 days after the third treatment
• Follow-up 2: the 24 hour period of the day 14 days after the third treatment

Other data will be sourced from:

• The initial visit questionnaires
• Follow-up questionnaires approximately 2 days after first treatment, 2 days after 2nd treatment, 2 days after final treatment, around day 21 and day 28.

Instruments

Data collection will be done through 8 main data collection instruments, drafts of which have now been developed:

• First contact questionnaire
• Crying Diary (including severity of other symptoms)
• Initial Visit Questionnaire - Part 1 (demographics and history)
• Initial Visit Questionnaire - Part 2 (success of blinding)
• Credibility and Expectancy Questionnaire (CEQ)
• Hospital Anxiety and Depression Scale
• Adjustments record
• Weekly telephone questionnaire

Power

The sample size is 160, calculated achieve 80% power, 5% significance level, and using an estimated standard deviation of around 2 (based on results from similar studies) for a 1 hour between-groups difference (identified as a 'clinically relevant reduction in crying time' from a parental survey) and allowing for a 25% attrition rate in the latter 2 weeks (based on experiences in similar studies).

Analysis:

Primary Outcomes

For the primary analyses comparisons will be carried out using all the data in one analysis which will assess the significance of the treatments, the blindness and the age groups in an ANCOVA. This will include interaction terms to assess whether the treatment effects are different for the different levels of blindness and for the different age groups and all baseline variables will be included as covariates.

A logistic regression analysis will be used to produce odds ratios for the comparison of treatment, blindness and age groups. The between-groups differences in proportion (risk ratio) of patients reporting 'complete resolution' of colic symptoms (PGIS) after first, mid-point and final treatments (primary outcome) and at follow ups (secondary outcomes) Adverse events will be reported with no analysis. Extraction of primary data from crying diaries will be performed by observers blinded to group allocation.

The primary analysis will be performed by a statistician blinded to group allocation.

Secondary Outcomes

The primary analysis of primary outcomes will be extended to days 21 and 28, as above.

The ANCOVA will be extended to evaluate the between-group differences in

• Cost of NHS resource usage
• Duration that the infant is fussy It is anticipated that parent compliance in completing the diary is likely to deteriorate during once the clinic visits have ceased. However, regular telephone calls will attempt to encourage continued completion. The primary study outcomes will be based on the first 15 days of the diary and it is anticipated that there will be greater compliance during this time.

Investigative Analysis

The exploratory investigations seek to establish any associations between the outcome measures (daily hours crying, presence of other symptoms, Parental Global Impressions of Severity (PGIS), Parental Global Impressions of Change (PGIC), and 'resolution' of colic) after first, mid-point and final treatments, and at follow-ups and any other predictor variables as gathered in the initial visit (Part 1). For example, :

• baseline daily hours crying
• initial symptom trajectory; worsening, plateau, improving
• baseline 'other symptoms' not at all, low, medium and high severity
• age at onset
• perinatal factors (Miller, 2007)
• treatment characteristics

Intention to treat v per protocol

Since this is a clinical study, the primary analysis will be done on an intention-to-treat basis, including any patients who fail to comply with the protocol for their group (e.g. dropping out of the treatment groups).

For some of the analyses, accurate data is required (e.g. actual adjustments performed) which cannot be imputed. In these cases, the analysis will be done on a per protocol basis.

Missing data

Previous studies have reported that the majority of dropouts occur in the no-treatment groups due to worsening symptoms and in the treatment groups due to resolution of symptoms, so data imputed at last value carried forward will represent a conservative approach. For the primary crying time outcome data, where there is a record for Day 8, this will be carried forward to Day 15.

Where reasons for withdrawal on the basis of worsening symptoms are given, this will be included as data points within the PGIS & PGIC analysis, so no imputation will be required.

Attempts will be made to contact any drop-outs to ascertain their reasons for choosing not to continue in the study and these will be reported separately.

For the other analyses, a longitudinal analysis of variance will assess changes over time for the treatment, blindness and age groups together with any interactions and covariates. Although there may be some loss of data in the later time periods, a repeated measures analysis should be able to cope with this without imputation of missing data.

Numbers of subjects responding to the study and progressing to each stage will be recorded on a CONSORT-style flow diagram.