The words “hospice”, “hostel”, and “hospital” share the same etymological root. In ancient times, the aged and infirm were often admitted to a hospice as a place to die. Typically the physical environment was spartan—walls bare except perhaps for a crucifix, and help limited to prayer. The spirit was otherwise ignored. Most, but not all, hospitals have overcome their monastic provenance and have evolved into cheerier environments. More and more hospitals now encourage the support of family and friends as an aid to recovery. But in too many adult hospitals, the physical environment still betrays its monastic unicellular roots. In health care, as in life, old traditions die hard.

The mere idea of a hospital dedicated exclusively to the care of children is a relatively recent concept. In the mid-19th century, Charles Dickens was a vigorous campaigner for the support of such a hospital in London—specifically, for The Hospital for Sick Children, Great Ormond Street. In Our Mutual Friend, published in 1868, he provided one of the first literary references to an exclusively paediatric hospital.¹ He describes how Mrs Boffin persuades an elderly woman to seek good care for an ill child:

“We want to move Johnny to a place where there are none but children; a place set up on purpose for sick children; where the good doctors and nurses pass their lives with children, talk to none but children, touch none but children and comfort and cure none but children.‘Is there really such a place?’ asked the old woman with a gaze of wonder.”

Dickens later describes how the child was permitted the comfort of bringing favourite toys to the hospital:

“At the Children’s Hospital, the gallant steed, the Noah’s Ark, the yellow bird and the officer in the Guards were made as welcome as their child owner.”

The early ancestors of the modern paediatric hospital (now, in many cases, wisely renamed health centres) included the foundling hospitals, best known for interminable stays, high mortality, and for being the sites where psychosocial deprivation, or “failure to thrive” was first identified as a clinical entity. The paediatric hospital’s ancestry also includes isolation hospitals for contagious disease (“fever hospitals”) and children’s orthopaedic hospitals often known, depressingly as (“hospitals for crippled children”).

Paediatrics has been a leader in fostering entirely new kinds of structural and functional health care facilities—places where families are not merely welcome, but play an increasing role as “care partners” in helping children recover. Hearing a child crying in a modern paediatric hospital is now the exception. Not so many years ago, it was the rule.

Hand in hand with the dramatic evolution of physical environments, and of our attitudes towards children’s family members (remember when they were referred to as “visitors”?) has been a remarkable trend towards minimising the number and duration of a paediatric hospital admissions without prejudicing children’s health and rate of recovery—often with the conviction that children’s health is better served through alternative approaches to paediatric care.

There are innumerable specific examples of this revolution in paediatric health care, including day surgery, home care, palliative care, short stay (observation) units, and most recently, tele-medicine to serve children living in remote areas.

A live tele-medicine consultation service has recently been shown to reduce dramatically the need to transfer critical care patients from a rural intensive care unit to a highly specialised central paediatric intensive care facility.²

In another innovative study of alternatives to traditional care, heart sounds from 87 patients with and without murmurs in distant rural areas of Norway were recorded with a sensor based stethoscope, e-mailed to a remote computer, and randomly distributed to four cardiologists, who had to categorise them as “no murmur”, “innocent murmur”, or “pathological murmur”. The cardiologists spent an average of 2.1 minutes per case. Mean sensitivity and specificity were 89.7% and 98.2% respectively, with low inter- and intra-observer variability.³ In a commentary that accompanied this report, Wren⁴ goes a step further, suggesting that the paediatric cardiologist him/herself could be replaced by a computer, citing a recent report showing 100% sensitivity and specificity of an artificial neural network in assessment of murmurs, a track record better than the average paediatric cardiologist!

If such innovations have achieved nothing else, they have certainly reduced dramatically the duration of paediatric hospital admissions, and many of us are convinced that, as a result, quality of care, speed of recovery, parent satisfaction, and other desirable objectives are being met. But if challenged, can we back up such fervent convictions with solid evidence?

In this issue, Dr David Ogilvie tackles this question head on.⁵ He challenges us to produce the evidence that hospital based alternatives to paediatric admission do more good than harm, and reports the findings of his systematic review of interventions designed to provide alternatives to acute admission in medical paediatrics. His search of the literature has been thorough, his inclusion criteria rigorous, and his evaluation evidence based. It is interesting to note that he unearthed only a single randomised controlled trial, and even this one was not without methodological blemishes. But despite such shortcomings, certain consistent themes emerge from Ogilvie’s analysis. For one thing, parent satisfaction levels with such alternatives to traditional care have been consistently high. Parents as a group are rarely wrong in their judgements, and one might question whether levels of parent satisfaction should be regarded as a subjective or an objective finding. The distinction may be rather arbitrary. An opinion, after all, is a kind of “fact”.

Dr Ogilvie puts us on notice to adopt more rigorous methods to evaluate any and all potential improvements in delivery of paediatric health care. He correctly advises us to perform our measurements using bi-directional scales, so that both positive and negative outcomes are assessed. There are, after all, few forms of treatment that do not include both positive and negative effects. It is our responsibility to establish for any health care innovation, that good outweighs harm by a wide, statistically significant margin.