The Science of Health

Is American Health care entering a big chill era?  This  question was posed in to-days Boston Globe.   Patients and physicians are concerned that computer screens wall them off from one another.  Questions in health care have traditionally sought answers in evidence.

 “To be master and servant within the hour                                                           That is the course of temporal power” 

Traditionally patients consulted with individual doctors of their choice.    The remit was clearly to diagnose illness, formulate a treatment plan, and advise accordingly.   The doctor was directly answerable to the patient.   All data collected during the consultation was confidential.   The entire system has undergone a transformation on an unprecedented scale.   The patient is now reviewed by a team, which includes doctors, nurses, allied health professionals and computer scientists.   Portions of this service are outsourced, particularly on the technology front.   The patient imparts information, the doctor is one of a number of people involved in collecting and collating this “big data”.   This in turn forms the basis for computational technologists who evaluate “unit performance”     All this data ultimately reverts to the Chief Financial Officer.   He can truly be described as the power behind the throne.

Patient & Doctor on same side of the cloud

Patient & Doctor on same side of the cloud

The escalation in health costs in Ireland is evidence of the extraordinary ability of new technologies to generate huge cost over-runs.   The impact is not confined to the financial sphere.   Numbers of patients on trolley increased from 0 to 500 while the cost of care quadrupled from 4bn euro to 16bn euro

Current issues in health care will ever be addressed until all stakeholders consent to view life from both sides of the cloud!



Primary Care in the 21st century: Make progress a reality

  Healthcare has interfaced with technology in the past decade. The fanfare has heralded a raft of new, sometimes exciting, sometimes bewildering tools. Individual health risk scores and protocols replace clinical judgement and care based on clinical examination and risk assessment.   These statistical scores mask often complex medical problems.

The parameters of any statistical analysis are optimised to maximise the predictive value to the patient.   However, in health care biological processes make individual variation more likely than in any other area of manufacturing or financial endeavour. Diagnostic accuracy is the cornerstone on which a healthcare system is constructed. Primary care physicians will always obtain as detailed a history as possible.   Experience teaches us to evaluate the information.   Laboratory and imaging play a key role in defining the precise nature of the illness.   A treatment plan is then formulated.

Evolutionary technology has delivered enormous benefits.   Implantable defibrillators, cardiac stents improved joint replacements, and robotic limbs immediately come to mind. The potential to leverage existing medical knowledge by the introduction of modern equipment and technology represents the next frontier in healthcare.  Reinstating the patient as the central figure, with the primary care physician enabled to access complementary imaging (eg ultrasounds),together with other diagnostic tools and specialist input offers real hope of faster, more precise diagnosis.

This vision also provides a mechanism for reintegration of care, in a system that has become extremely fragmented.   Rapid diagnosis and consensus treatment plans for challenging patients with complex multi-system problems  would also be facilitated.

#EHR : Electronic Health Records – what information should they contain


Health insurer Premera hit by “sophisticated cyberattack ……/99416_healthinsurerpremerahit-by-sophistica

3 hours ago – Cyber criminals may have gained access to members’ Social Security numbers, addresses and more.

Electronic health records make sense. Compiling an accessible record of health statistics with vaccination records, blood group etc can make the difference between life and death in the event of a serious road traffic accident or other critical event.   Duplication of investigations is avoided.   Small changes in blood count can alert the possibility of early colon cancer.
The question which must be asked is ‘How much data should your electronic health record contain?’  Should health records be aligned with social security numbers?  Would a unique medical identifier make the system more secure?
Never in the history of human endeavour  has so much personal information been available on so many to so few.   To-night serves as yet another reminder of the vulnerability of these data bases to cyber attack.   It is also noteworthy that insurance companies can be sold on.    Portions of the work can be outsourced.   Data bases may be used to facilitate data mining.
Confidentiality is an unwritten contract between patient and doctor.    It defines the relationship of trust and mutual respect.   Accurate history taking is facilitated.    The knowledge that this information may then be stored electronically and transferred to a third-party will inevitably profoundly alter confidence in the system.
A discussion on how best to marry the benefits of electronic health records with the subtext of preserving patient confidentiality is now mandatory.

It ain’t what you do, it’s the way that you do it

Thirty years ago to-day the first .com name – was registered.  It heralded a new era in human endeavour.   The outcome is truly transformative.   Thousands of domain names have now been registered.   Billions of euros worth of trade is conducted online.   Information is readily accessible with the click of a mouse.

Those who design computer systems are justifiably proud of their achievements.   Electron microscopy has enabled cell structure visualisation to an extent not previously considered possible.  Big data has facilitated demographic knowledge of unprecedented proportions.

The challenge now is to incorporate this information in a meaningful way.   It is vitally important, particularly in health care, to recognise that computers are only as effective as the information input.   To express this in colloquial terms ‘Rubbish in, rubbish out’   Diagnostic accuracy is a cornerstone of medicine.   A diagnosis defines an illness.   In healthcare, computerization and digitization has allowed storage and retrieval of data.   It has facilitated the identification of single cancer cells using increasingly fine biopsy needles.  It is now mandatory that this level of detection is married with conventional wisdom and knowledge of disease.   The human cost of treatment for cancer must  be set against the desire for early detection.

The cost increases and illness levels witnessed when computerised mechanisms were introduced is testimony to the net health effects of excessive zeal and over-reliance on Breakdown of revenue funding

technology.   The increase in incidence of diagnosed illness will inevitably be reflected in increased pressure on hospital services, increase illness rates and an increase in health costs, but without any absolute improvement in wellbeing or mortality as assessed on a population basis.

Increased awareness of the role of technology in health care together with an increased engagement by clinicians in the configuration of equipment combined with a holistic approach to health care offers the prospect of a brighter healthier future for all


#Cancer and Technology


, , , , ,

Can an algorithm tell if you have cancer?  This question captures the complexity of unproven digital technology in health care,.    It defines the hopes  of a generation of healthcare technology workers and the skepticism of traditional physicians.

Physicians have traditionally employed history from the patient,with clinical examination and laboratory findings to make a diagnosis.   Novel cancer screening methods omit the first two steps.   Clients are invited for screening based on age gender or propensity for disease.   Symptoms of illness are not a requirement.   Cancer is diagnosed using one modality only: laboratory and radiological findings.

Design choices have an enormous impact on a mechanised diagnostic system.   Digital images can be manipulated after acquisition.   Image enhancement can contribute to an excess of illness diagnosed.   Basic questions must now be addressed.   Actual clinical findings must be considered the gold standard against which diagnostic imaging is measured.   The correlation coefficient rather than standardised detection ratios or algorithms should determine whether new technology is fit for purpose.   Clinicians preferably with actuarial assistance would determine the suitability of new technology for the clinical setting.

Early diagnosis is recognised to confer superior outcome for cancer patients.   The objective of every clinician is to detect cancer at Stage 1 when cure rates are almost guaranteed.   Patients also like the reassurance that, although they have cancer, it was detected early because of screening.   Both factors result in a bias towards acceptance of a cancer diagnosis.  Some particularly wise surgeons have insisted on a second opinion on the pathology to ensure that cancer is actually present.   This commendable practice should be encouraged and become standard.

Accurate correlation would ensure that screening did not result in increase in incidence of cancer.  It would ensure that no-one was treated for cancer in the absence of illness.    New technologies open vast new horizons in health care.   Combining traditional physician knowledge and wisdom with the unimaginable realm of petabytes and exabyte gathered at terahertz speeds has the potential to deliver a stellar quality of care

Artificial intelligence and #Healthcare


, ,

‘The age of artificial intelligence is finally at hand.  Will we master it or will it master us?’   This question was addressed in the New York Times by David Brook last week. The issue of computerization and it’s integration into mainstream living is arguably more pertinent to health care than in any other domain.

Atul Gawande discussed the challenges faced by doctors and health staff who deliver front line services to patients in a very interesting Washington Post blog

Atul Gawande on the ultimate end game – The Washington Post
Two very disparate services are described.   One portion of the budget is allocated to treat as many patients as speedily and effectively as possible.   The other portion is allocated to fee-for-service, where the demand is to accomplish the maximum number of procedures in the shortest possible time.
The breast cancer screening programme affords an insight into managerial input and the process by which some of our health care is delivered.   New digital technologies are introduced.   Unit performance is assessed based on the number of cancers detected.   Algorithms are employed to derive images.   Patient safety is not a priority.  The inevitable consequence of this approach is an increase in the number of false positive readings.   Unfortunately, a large number of cancers are also missed or caused during the screening process.   Computerisation of pathology reporting, again with a statistical and engineering background, would emphasise the role of positive predictive values(PPV) in system evaluation.   This creates a duplicate problem.   More biopsies result in more cancers.   This in turn validates the requirement for biopsy.
Atul Gawande has averted to the conundrum.  He states that engineers will have the power to build the frameworks through which the rest of us make decisions and to steer our choices.    Engineers who design health care technology must become more cognisant of the requirement to prioritise patient safety.   It is self-evident that twice as many breast cancers are twice as bad for those diagnosed and for the economy as a whole.   Parameters for system evaluation must be clarified.
Knowledgable experts such as Dr Gawande must be engaged and included. The systems currently operational have seen excellent clinicians despair and shun leadership roles. A comprehensive evaluation of new systems and outcomes is urgently required. Correlation between physical findings and computer generated images is vital.   Standardised detection rates are less important than interval cancers in assessing the efficacy of a screening programme.    it is also true that a screening programme should not inherently increase the incidence of an illness.   Any initial round screening may detect illness earlier, but this increase should be balanced out over time.
New technology poses novel opportunities and many new challenges.   Doctors are now outnumbered by management in the health care sector.   Decisions are taken and validated by financial managers.   Delivery of a good health care system in the presence of inaccurate or dangerous equipment is as difficult for a doctor as baking a cake using cement for a cook.   Opportunities for improvement abound.  Lets take them.




Tackling childhood #obesity at source

Pregnancy clip

Pregnant: Eat smart Eat healthy Avoid alcohol Target: No more than 28lb weight gain during pregnancy

Childhood obesity regular makes headlines, with 25% of children in Ireland considered overweight and 5% frankly obese.

For the first 9 months of a child’s life, the baby is dependent solely on the nutrition from the mother.   Foods, alcohol and drugs consumed by the mother are passed across the placenta and shared with the baby.  Mothers have a very special role to play in their child’s weight

Alcohol passed to the baby results in a baby born with foetal alcohol syndrome.   The baby has stunted growth, &identifiable facial characteristics.   More significantly, brain growth and development are affected, resulting in poor memory, attention deficit etc. This can be checked on Foetal alcohol spectrum disorder (FADS) hhtp://

Foods which are known to contribute to excessive weight gain in adults will also contribute to weight gain in the unborn baby. The sugars from carbonated drinks and snack foods will cross the placenta.   Ideally babies should weigh no more than 4kg or 8lb7oz at birth.   Babies who weigh more than 4kg are said to have macrosomia.    The delivery is more difficult.  Caesarean section rate increases.   Baby is more prone to develop problems with blood sugar, affecting brain development.   These babies are more likely to develop diabetes, the metabolic syndrome and breast cancer as adults.

Rule of thumb: Feed yourself and your baby during pregnancy with as much care as you would feed it when born. Limit the consumption of carbonated drinks, snack foods, fries chips and alcohol. 

Make a healthy body, a healthy mother & a healthy baby your priority!  Nobody but you can feed the baby for the first & most important formative nine months of your child’s life.,

#mammograms Who in the world are they good for?


, ,

Early detection of cancer is very advantageous.   The awareness created by screening has undoubtedly contributed significantly to earlier presentation of cancers and reduced mortality.

Overdiagnosis constitutes a very serious consequence of screening.  Cancers are detected on X-ray which are not palpable.   These identified cancers would likely never progress to significant disease in 10% of patients.   These women progress to mastectomy with approximately 16 sessions of chemotherapy and possibly radiotherapy.  This treatment in and of itself has considerable side effects and a substantial impact on health.  BMJ 2006;332;689-692; originally published online 3 Mar 2006;  This fact is rarely alluded to by those who have been diagnosed and consider themselves fortunate to have the cancer detected early.

False positive readings i.e. recall for further evaluation is anticipated in 40% of women routeinly screened  biannually over a 10 year period.   This causes unnecessary anxiety and may even result in  biopsy.

Interval cancers constitute a further headache for those involved in screening.  The interval cancer rate is recorded at 20% to 38% of expected incidence.   The interval cancers are generally large aggressive cancers.   They are rarely Stage 1 and may even be stage 4 cancers.   The question as to whether these cancers are missed or caused by the compression/X-ray delivery to sensitive breast tissue is one which requires further consideration.   Women should not be misled into thinking that screening mammography guarantees any cancer will be stage 1.

Does screening affect mortality?  The answer to this is possibly not.   Screening was introduced to Southern Ireland 15 years after it’s introduction to Northern Ireland.   The mortality in both juristictions reduced at a similar rate.  Curiously, a blip in the downward mortality trend was witnessed in both areas at the time of introduction of mammography screening.   BMJ. 2011 Jul 28;343:d4411. doi: 10.1136/bmj.d4411

Mortality outcomes with & without screening

Mortality outcomes with & without screening

This data validated by #WHO graphically illustrates the positive impact of awareness and comprehensive care structures on breast cancer mortality.   It does little to justify the burden on health and healthcare of screening.

Conclusions of Swiss Medical Board on mammography screening  Those scientists and epidemiologists who have studied mammography in detail recognise that the advantages conferred by early detection are frequently surpassed by the harms caused.Conclusions of Swiss Medical Board on mammography screening

The position adopted by the Swiss Medical Board seems most prudent in light of the above scientific data.