Ultimately, assessments of needs, either met, unmet or overmet, further explored by service researchers like Bobevski et al. (Reference Bobevski, Rosen and Meadows2017), should inform policy analysts to come to population-based estimates about the number of people in need, the types of interventions and services required, as well as the length and intensity of these services. An example can be found in the recent Canadian policy document ‘strengthening the case for investment’. This document, released months before the federal government announced a 10-year mental health transition fund in its March 2017 budget (Table 4; MHCC, 2017; see also Lesage et al. Reference Lesage, Bland, Musgrave, Jonsson, Kirby and Vasiliadis2017), implies the best level of service provision, not only in treatments provided by specialists and primary care levels, but also for prevention and promotion interventions by public health initiatives.

Bobevski et al. (Reference Bobevski, Rosen and Meadows2017) have well recognised the limitations of transversal epidemiological surveys in assessing a basic dimension of needs, like the presence of a mental disorder. No survey instrument covers all disorders, with personality disorders or childhood disorders extending into adulthood like attention-deficit/hyperactivity disorder being important examples. Mental disorders are chronic diseases and the most common ones, like anxio-depressive disorders, are so highly co-morbid with other common chronic diseases (i.e. diabetes, hypertension, cardio-vascular diseases, asthma), that they represent more of systemic disease rather than separate diseases (Lesage, Reference Lesage2015). Health care needs, including needs for specific treatments like antidepressants and psychotherapy, should not be assessed in a primary care context separately from other chronic diseases the patient and family physician encounter. They should be addressed simultaneously and over the lifetime. Precaution should be exercised to avoid giving priority to populational assessments to plan services, the top–down approach; while health care needs can only be established by a dialectic of individual and populational perspectives.

Efforts to introduce standardised assessments of needs, in epidemiological surveys, started with the development of a standardised assessment of needs for the severely mentally ill by the MRC Social Psychiatry Unit in the 1980s (Brewin et al. Reference Brewin, Wing, Mangen, Brugha and MacCarthy1987). Needs were defined in terms of the type of impairment, the factors causing social disablement, the model of treatment or other interventions required to meet it. If an individual is socially disabled, in association with a mental disorder for which an effective and acceptable form or model of care exists, either for amelioration or prevention, the individual needs that intervention (Wing et al. Reference Wing, Brewin, Thornicroft, Thornicroft, Brewin and Wing1992). The first step of the three-step procedure was an interview with the patient to provide systematic assessment of the problems using a standardised measurement of diagnosis and of incapacity. Afterwards, service receipts were established through a standardised interview with patient and, when available, a review of case records. Lastly, a panel generally composed of a psychiatrist, a psychologist and sometimes a social worker, reviewed a narrative of the problems, of the trajectory of services utilisation (interventions and settings) and of the interviewer's impressions, to rate a systematic instrument recording for various clinical and social areas, the problems, the required interventions. By comparing required interventions with recent interventions, the instrument allowed to ascertain if the needs for care were met, unmet, overmet or impossible to meet. Instruments were then developed to capture the perception of needs for care of the clinical team or the patient, via interviews rather than the judgement of a panel. Differences in problem assessments, especially for priorities, were found when comparing the staff's assessment of need to the patient's (Slade et al. Reference Slade, Phelan, Thornicroft and Parkman1996; Comtois et al. Reference Comtois, Morin, Lesage, Lalonde, Likavcanova and L’Écuyer1998). In the 1990s, there was the development of a briefer instrument applicable to common disorder cases seen in the community. With its seven items, the MRC Community version of the Needs for Care still required a panel's judgement to be systematically recorded (Bebbington et al. Reference Bebbington, Brewin, Marsden and Lesage1996). Finally, there was the development of the interviewer, or self-administered Perceived Needs for Care Questionnaire (PNCQ) in Australia (Meadows et al. Reference Meadows, Harvey, Fossey and Burgess2000).

A first generation of epidemiological surveys assessed needs by combining standardised instruments assessing (i) mental disorders; (ii) incapacity; (iii) distress; (iv) services utilisation. For example, Shapiro's trial (Shapiro et al. Reference Shapiro, Skinner, Kramer, Steinwachs and Regier1985) considered in need, either (1) prior 6 months use of services for mental health problem; or (2) presence of two or more manifestations of (2.1) one or more DIS disorders present in the last 6 months; (2.2) general health questionnaire score of 4 or more (current symptoms); (2.3) respondent's report of having been unable to carry out usual activities in the past 3 months for at least 1 day. They found that 7.1% of the population had mental health visits in the past 6 months, while 6.7% were not in contact but met two of the previous criteria, resulting in a total of 13.8% of the population in need for care. A second generation included systematic needs assessment procedures. Bebbington et al. (Reference Bebbington, Marsden and Brewin1997) applied the MRC community needs for care assessment in a two-stage survey, with 408 subjects in the second stage. They found 9.8% of 1-month International Classification of Diseases, 10th Edition (ICD-10) mental disorders prevalence and 9.5% of the population in need for treatment according to the needs for care assessment schedule-community version (NFCAS-C) procedure; 50% of needs were met. Applying a similar design and instrument in Montreal, but with a panel including a general practitioner in addition to psychologists, Fournier et al. (Reference Fournier, Lemoine, Poulin and Poirier2002) found a higher prevalence of needs (18%). While two-thirds of the needs were unmet, 17% of the needs were met, 11% were impossible to meet and no needs were overmet. Finally, the most recent Statistics Canada mental health and well-being survey among adults applied the PNCQ and found that 18% of the Canadian population had a need for care; about 10% of the population expressed a need for medication and 12% for talk therapy (more than one type of need can be expressed). However, if 90% of the needs for medication were met, 65% of the needs for psychotherapy were met (Sunderland & Finlay, Reference Sunderland and Finlay2013).

The analysis conducted by Bobevski et al. (Reference Bobevski, Rosen and Meadows2017) relied on data from the first generation of populational surveys and perspective on needs. It did not include the perspective of patients, nor of the clinical judgements of a panel of clinicians around a case history. The latter is however open to variations according to the composition of the panel, for example, including general practitioners or psychologists with different psychotherapy orientations (i.e. dynamic v. cognitive behavioural). It is of interest that the global estimate of needs in Canada, either from the Montreal survey using the panel approach of the MRC community NFCAS, or from the population survey using the respondents’ perception of needs with the PNCQ, were both of 17–18% of the population, thus reinforcing the idea of a socio-cultural construction of mental disorders and needs for mental health care.

Other contextually based individual assessments of needs that can inform programmes and systems are represented by systematic audits of outcomes, such as suicide. In the UK, the national confidential enquiry on suicide by people with mental disorders investigate, with a 40-page questionnaire, the unmet needs for care and services in the months preceding death among patients in contact with specialist services (about 25–50% of all suicide). Local and regional authorities following the recommendations based on aggregation of individual audits demonstrated lowering rates of suicide (While et al. Reference While, Bickley, Roscoe, Windfuhr, Rahman, Shaw, Appleby and Kapur2012). A systematic audit of all suicide cases in one Canadian province made use of the last generation of individual needs for care and services assessment and allowed to differentiate unmet needs for specific treatment from care management in specific treatment settings (i.e. in the Emergency Department; in the primary care provider's office). This audit also found unmet needs for promotion or prevention interventions at the system level (i.e. mental health literacy programme in the school – see Lesage et al. Reference Lesage, Seguin, Guy, Daigle, Bayle, Chawky, Tremblay and Turecki2008; Wasserman et al. Reference Wasserman, Hoven, Wasserman, Wall, Eisenberg, Hadlaczky, Kelleher, Sarchiapone, Apter, Balazs, Bobes, Brunner, Corcoran, Cosman, Guillemin, Haring, Iosue, Kaess, Kahn, Keeley, Musa, Nemes, Postuvan, Saiz, Reiter-Theil, Varnik, Varnik and Carli2015). If such audits provide, by their recommendations areas of priorities, the amplitude, cost, benefits and budget impact of these recommendations can only be assessed by modelisation (see e.g., Vasiliadis et al. Reference Vasiliadis, Lesage, Latimer and Seguin2015).